Ecology of Culicoides Biting Midges and Their Role As Vectors of Epizootic Hemorrhagic Disease Virus on Florida Big Game Preserves

ECOLOGY OF CULICOIDES BITING MIDGES AND THEIR ROLE AS VECTORS OF EPIZOOTIC HEMORRHAGIC DISEASE VIRUS ON FLORIDA BIG GAME PRESERVES

BETHANY L MCGREGOR

A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY

UNIVERSITY OF FLORIDA

2019

To my family, friends, and mentors who have always supported and encouraged me

ACKNOWLEDGEMENTS

I thank my family for their encouragement, support, and patience throughout my graduate education. I thank my friends for offering humor and companionship during the most stressful of times. I thank my lab mates for serving as a source of inspiration and a sounding board for ideas. I thank the FMEL lab technicians for their unending help towards completion of this research. Finally, I thank my committee and mentors for their guidance and advice throughout the PhD process.

TABLE OF CONTENTS page

ACKNOWLEDGEMENTS ...... 4

LIST OF TABLES ...... 8

LIST OF FIGURES ...... 10

LIST OF ABBREVIATIONS ...... 11

ABSTRACT ...... 12 CHAPTER

1 LITERATURE REVIEW ...... 14

2 HOST USE PATTERNS OF CULICOIDES SPP. BITING MIDGES AT A BIG GAME PRESERVE IN FLORIDA, USA ...... 25

Methods ...... 29 Field Methods ...... 29 Laboratory Methods ...... 31 Data Analysis ...... 33 Results ...... 34 Forage Ratios ...... 34 Habitat Associated Host Use ...... 36 White-tailed Deer Aspiration ...... 37 Discussion ...... 38

3 VERTICAL STRATIFICATION OF CULICOIDES BITING MIDGES AT A FLORIDA BIG GAME PRESERVE ...... 52

Methods ...... 54 Field Collections ...... 54 Laboratory Methods ...... 56 Data Analysis ...... 58 Results ...... 58 Total Collections ...... 58 Ground versus Canopy Use by Abundant Species ...... 59 Funnel Suction Traps ...... 60 Habitat Associations of Abundant Species ...... 60 Physiological Status ...... 61 Blood Meal Analysis ...... 61 Discussion ...... 63

4 IMPLICATING CULICOIDES STELLIFER AND CULICOIDES VENUSTUS (DIPTERA: CERATOPOGONIDAE) AS VECTORS OF EPIZOOTIC HEMORRHAGIC DISEASE VIRUS ...... 79

Methods ...... 82 Culicoides Sampling and Virus Detection During EHDV Epizootic in Northern Florida ...... 82 White-tailed Deer Sampling and Virus Detection during the 2017 Epizootic ..... 84 Statistical Analysis of Virus Data ...... 85 Live Animal Aspiration ...... 86 Blood Meal Analysis ...... 86 Seasonal abundance of C. stellifer and C. venustus ...... 87 Results ...... 88 Virus Detection in Field Collected Midges during an EHDV Epizootic ...... 88 Animal Mortality due to EHDV ...... 89 Live Animal Aspiration ...... 91 Blood Meal Analysis ...... 91 Seasonal Abundance of C. stellifer and C. venustus ...... 92 Discussion ...... 92

5 VECTOR COMPETENCE OF CULICOIDES SONORENSIS FOR EPIZOOTIC HEMORRHAGIC DISEASE VIRUS SEROTYPE 2 STRAINS FROM CANADA AND FLORIDA ...... 106

Methods ...... 108 Culicoides sonorensis colony ...... 108 Viral Strains ...... 109 Bloodfeeding ...... 110 Sample Collection ...... 110 Honey Cards ...... 111 Sample processing ...... 112 Statistical analysis ...... 113 Results ...... 114 Infection, Dissemination, and Transmission Rates ...... 114 Viral Titer Analysis ...... 115 Honey Card Versus Capillary Assay Comparison ...... 116 Discussion ...... 117

6 VECTOR COMPETENCE OF CULICOIDES INSIGNIS (DIPTERA: CERATOPOGONIDAE) FOR EPIZOOTIC HEMORRHAGIC DISEASE VIRUS SEROTYPE 2 ...... 127

Methods ...... 130 Blood Feeding Trial Specimen Collection...... 130 Blood Feeding Apparatus Design ...... 130

Blood Feeding Trial Variables ...... 131 Infection Trial Culicoides insignis Collections ...... 131 Viral Blood Feeding ...... 132 Sample Processing ...... 132 RNA Extraction and Virus Detection ...... 133 Statistical Analyses ...... 134 Intrathoracic Inoculation ...... 135 Virus Screening of Field-Collected C. insignis ...... 135 Results ...... 136 Blood Feeding Trials ...... 136 Infection, Dissemination, and Transmission ...... 136 Microinjection ...... 139 Virus Screening of Field-Collected Midges...... 139 Discussion ...... 139

LIST OF REFERENCES ...... 148

BIOGRAPHICAL SKETCH ...... 173

LIST OF TABLES

Table page

2-1 Total big game abundance estimates on the Gadsden County, Florida big game preserve during the 2015 and 2016 trapping season ...... 46 2-2 Total bloodmeals taken by all Culicoides spp. on game mammals, non- game mammals, birds, and other sources on the study preserve in 2015 (July-December) and 2016 (January-December)...... 46 2-3 Non-game mammalian and avian blood meals taken by Culicoides spp. during the full study period (July 2015-December 2016) ...... 48 2-4 Total game mammal (Cervidae and Bovidae) blood meals taken by Culicoides for which >5 game mammal blood meals were recovered ...... 49 2-5 Results of linear regression testing for differences in host use (proportion of blood meals from each host species) among major habitat types ...... 50 3-1 Total Culicoides collected per species and height during the 2016 and 2017 sampling period ...... 74 3-2 Pearson’s chi-square results for distributions of physiological status in ground and canopy traps ...... 76 3-3 Physiological status distribution for ground and canopy collected Culicoides in 2016 and 2017 ...... 78 4-1 Summary of sampling locations, sampling frequency, and EHDV-positive samples collected from dead deer in 2017 ...... 100 4-2 Total Culicoides sampled, number of pools per species, and EHDV positive pools during the EHDV epizootic in northern Florida from August-October, 2017 ...... 101 4-3 Serotypes of EHDV recovered from C. stellifer, C. venustus, and O. virginianus during the 2017 outbreak ...... 103 5-1 Daily infection (body), dissemination (legs), and transmission (saliva) rates for the Can-Alberta strain and the Florida strain of EHDV-2 in Culicoides sonorensis ...... 122 6-1 Culicoides insignis blood feeding trial membrane and blood combinations ..... 145 6-2 Pairwise Wilcoxon rank sum tests on blood-feeding rates by membrane type . 146

6-3 Culicoides insignis blood-feeding trial starvation periods ...... 146 6-4 Summary of the Culicoides insignis EHDV vector competence trials ...... 147

LIST OF FIGURES

Figure page

2-1 Map of the big game preserve located in Gadsden County, FL and Culicoides spp. by habitat type ...... 45 2-2 Forage ratios for ungulate species and family by Culicoides debilipalpis and Culicoides stellifer ...... 47 2-3 Comparison of abundance of 10 Culicoides spp. in total trap counts of all physiological statuses, bloodmeals on white-tailed deer, and aspirations from white-tailed deer ...... 51 3-1 Map depicting property boundaries, habitat classes, and trap sites at a big game farm ...... 72 3-2 Funnel suction traps, biting midges captured and their physiological status ...... 73 3-3 Habitat associations and vertical stratification of Culicoides species ...... 75 3-4 Vertical distribution, physiological status and host use of Culicoides ...... 77 4-1 EHDV infection rate of Culicoides at five Florida deer farms during HD epizootic, 2017 ...... 102 4-2 Host use of Culicoides stellifer and Culicoides venustus at the Gadsden-1 site ...... 104 4-3 Seasonality of C. stellifer, C. venustus, and EHDV related mortality at the Gadsden-1 farm in 2016 ...... 105 5-1 Viral titers in the bodies, legs, and saliva of Culicoides sonorensis infected with the Can-Alberta or Florida strain of EHDV-2...... 123 5-2 Viral titer relationships between bodies, legs and saliva of C. sonorensis infected with two strains of EHDV-2 ...... 124 5-3 Virus detection by saliva collection method for two strains of EHDV-2 in Culicoides sonorensis...... 125 5-4 Honey card and capillary assay titers for two EHDV strains in Culicoides sonorensis over time...... 126 6-1 Blood-feeding apparatus used for blood-feeding trials and C. insignis vector competence trials ...... 145 6-2 Viral titers in bodies, legs, and saliva for two populations of C. insignis at two inoculation titers each ...... 147

LIST OF ABBREVIATIONS

AHSV African Horse Sickness Virus BTV Bluetongue Virus CDC Centers for Disease Control and Prevention DPF Days Post Feeding EHDV Epizootic Hemorrhagic Disease Virus HD Hemorrhagic Disease, used when describing the combined diseases caused by both bluetongue virus and epizootic hemorrhagic disease virus MLE Maximum Likelihood Estimate for estimating infection rates based on variable pool sizes PCR Polymerase Chain Reaction PFU Plaque Forming Units, the number of infectious particles present in viral solution qRT-PCR Quantitative Reverse-Transcription Polymerase Chain Reaction

TCID50 Median Tissue Culture Infective Dose, the dose at which half of the inoculated tissue culture is infected WNV West Nile virus

Abstract of Dissertation Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy

ECOLOGY OF CULICOIDES BITING MIDGES AND THEIR ROLE AS VECTORS OF EPIZOOTIC HEMORRHAGIC DISEASE VIRUS ON FLORIDA BIG GAME PRESERVES

Bethany L McGregor

May 2019

Chair: Nathan D. Burkett-Cadena Major: Entomology and Nematology

Culicoides Latreille biting midges (Diptera: Ceratopogonidae) are vectors of over

50 arboviruses worldwide. Epizootic hemorrhagic disease virus (EDHV) is one

Culicoides-borne arbovirus that causes morbidity and mortality to white-tailed deer.

White-tailed deer farming is a growing industry in the United States, necessitating research on EHDV epidemiology to find ways to reduce negative impacts to farmed deer. This six-chapter dissertation investigated Culicoides ecology on deer farms and the role of Culicoides as virus vectors in Florida. Chapter 1 is a review of the literature on Culicoides biology and ecology. Chapter 2 investigated host use of big game animals on game farms, finding that Culicoides stellifer and Culicoides debilipalpis fed preferentially on EHDV host animals including white-tailed deer, fallow deer, and elk.

Furthermore, there was an overall avoidance of bovid species and a preference for cervids. Chapter 3 investigated vertical stratification of midges, finding a significant association with the canopy for C. arboricola, C. biguttatus, C. debilipalpis, C. haematopotus, C. insignis, and C. stellifer. Greater abundance of blood-engorged

12 midges was found in the canopy compared with ground-level; however, 98% of blood meals analyzed from canopy-collected midges were from ground dwelling mammals. In

Chapter 4, qRT-PCR was used to detect virus positive Culicoides pools during an active

EHDV outbreak resulting in 20 positive pools: six from C. stellifer and 14 from C. venustus. Chapter 5 investigated the differences in vector competence of a confirmed

EHDV vector, C. sonorensis, for two strains of EHDV-2: a strain isolated in 2016 in

Florida and a strain isolated in Alberta, Canada in 1962. Overall, infection and dissemination rates were higher for the Florida strain while the viral titers in bodies, legs, and saliva were higher for the Alberta, Canada strain. Finally, Chapter 6 investigated the vector competence of field-collected C. insignis for EHDV-2. At high titers, C. insignis is a competent vector for EHDV-2 supporting replication of the virus in bodies, legs, and saliva of infected midges. At lower titers, the virus was blocked by infection barriers within the midges. Overall these experiments increase our knowledge on Culicoides ecology and vectors of EHDV in Florida.

CHAPTER 1 LITERATURE REVIEW

Culicoides biting midges are small (1-3mm) hematophagous flies in the order

Diptera, family Ceratopogonidae. There are currently more than 1,500 described species of Culicoides worldwide, with 150 present in North America, and 49 in the state of Florida (Blanton and Wirth 1979, Borkent and Grogan 2009, Borkent 2016).

Morphological identification of Culicoides species is largely based on wing patterns, leg banding patterns, characteristics of the sensory pit on the 3rd palpal segment, and number of oothecae (Blanton and Wirth 1979). Due to cryptic species and challenges with identification of juvenile specimens, increasing efforts are being made to establish barcoding libraries for Culicoides around the world (Ander et al. 2013, Harrup et al.

2016, Talavera et al. 2017, Bakhoum et al. 2018).

Culicoides exhibit a typical dipteran life cycle with an egg, larva (four instars), pupa, and adult stage. The eggs of Culicoides are cigar-shaped and range in size based on species. For a large species such as Culicoides sonorensis Wirth and Jones, eggs are around 432-477µm in length and 63µm in width (Abubekerov and Mullens

2017). Smaller species, such as Culicoides oxystoma Kieffer, have eggs ranging from

290-302µm in length and mean width of 41µm (Harsha et al. 2017a). Eggs are covered in small projections of various shapes and sizes called ansulae (Abubekerov and

Mullens 2017, Harsha et al. 2017a). The purpose of the ansulae is currently unknown, although there is speculation that these structures may act as a plastron to encourage respiration in low oxygen environments or to attach the eggs to the substrate (Becker

1960, Campbell and Kettle 1975, Wong et al. 2018). The length of the egg stage is variable depending on species, but typically ranges between 2-5 days (Sun 1974,

Harsha and Mazumdar 2015, Barceló and Miranda 2018). While most species are believed to be unable to survive drought conditions, there is some evidence of drought tolerance in the eggs of colony-reared C. sonorensis (McDermott and Mullens 2014).

Additionally, the eggs of C. sonorensis are fairly cold-tolerant, able to withstand short periods at temperatures as low as -20˚C (McDermott et al. 2017).

Culicoides larvae possess a heavily sclerotized head capsule (Sun 1974,

Abubekerov and Mullens 2017), and a body that is transparent during early instars, progressing to a creamy white in later instars (Linley and Kettle 1964, Sun 1974). The thorax of some species also possess patches of pigmentation, which are often used in their identification (Kettle and Lawson 1952, Linley and Kettle 1964). The 1st instar larvae of some species possess a proleg, however this is not a conserved trait across the genus and is lacking in many species (de Meillon and Wirth 1991, Abubekerov and

Mullens 2017). All larval instars possess anal papillae that can be everted but are often held within the body (Murphree and Mullen 1991, Abubekerov and Mullens 2017). Anal papillae are used for the exchange of chloride ions with the environment and, therefore, serve an important osmoregulatory purpose (Reeves 2008). The length of the larval stage in Culicoides is extremely variable. At cold temperatures, some Culicoides 4th instar larvae can enter diapause and remain in this stage for six months (Dove et al.

1932). Typically, this stage lasts 1-4 weeks, depending on temperature, diet, and sex

(Linley 1969, Mullens and Rutz 1983, Veronesi et al. 2009, Harsha et al. 2017b).

The pupae of Culicoides are similarly colored to the larvae ranging from yellow to dark brown (Blanton and Wirth 1979). This life stage is characterized by having a cephalothorax bearing respiratory horns for respiration at the substrate surface (Linley

15 and Kettle 1964, Harsha and Mazumdar 2015). Pupal Culicoides can move throughout water and moist semi-solid substrate using abdominal movement and are capable of stratifying vertically within the substrate, sometimes moving to depths of 10cm (Dyce and Murray 1966, Uslu and Dik 2006). This life stage is typically short at just two to three days, although it can last up to a month depending on factors such as temperature and species (Mellor et al. 2000).

The juvenile stages of Culicoides generally inhabit semi-aquatic habitats, although these habitats are extremely varied depending on species. Examples of larval habitats include pond and stream edges, tree-holes, muddy puddles, animal footprints, sandy beaches, mangrove swamps, and animal manure (Jones 1961, Jamnback 1965,

Blanton and Wirth 1979). A variety of factors relating to soil chemistry, salinity, and organic content influence appropriate conditions for larval habitation; however, ranges of these factors are typically species specific (Battle and Turner 1972, Schmidtmann et al. 1996, Uslu and Dik 2010). Within these diverse habitats, juvenile Culicoides diets are variable between different species. Culicoides arakawae and C. schultzei were found to be extremely predaceous and cannibalistic in laboratory studies of established colonies

(Sun 1974). Culicoides furens were found to feed preferentially on green algae, switching to predation only in the absence of this preferred food source (Aussel and

Linley 1994). Culicoides sonorensis is a generalist, feeding opportunistically on a diverse array of microorganisms and nematodes when provided (Jones et al. 1969,

Mullens and Velten 1994).

Adult Culicoides are varied in appearance, ranging in color from bright yellow to brown to gray. Most species have unique wing patterns that aid in differentiating

16 between species morphologically (Blanton and Wirth 1979). Some adult Culicoides females are autogenous and receive adequate nutrition from the larval stage in order to produce their first batch of eggs (Koch and Axtell 1978, Linley 1983). However, most female Culicoides are anautogenous, requiring blood from a vertebrate host in order to develop their first and subsequent batches of eggs (Blanton and Wirth 1979).

Understanding aspects of the biology and ecology of adult Culicoides is important as this genus is responsible for transmission of greater than 50 arboviruses worldwide (Mellor et al. 2000). While some Culicoides species do transmit human pathogens, including Oropouche virus and multiple filarial nematodes associated with mansonellosis (Mazumdar and Mazumdar 2016), Culicoides are primarily known for their role in transmitting viruses of veterinary importance (Mellor et al. 2000). Some of the most significant viruses transmitted by Culicoides are in the family Reoviridae, genus Orbivirus including African horse sickness virus (AHSV), bluetongue virus (BTV), and epizootic hemorrhagic disease virus (EHDV). All three of these pathogens can lead to extreme morbidity and mortality in vertebrate host animals (Mellor and Hamblin 2004,

Szmaragd et al. 2007, Savini et al. 2011). Combined, BTV and EHDV are often collectively termed hemorrhagic disease (HD) due to their similar pathologies (Smith et al. 1996b, Stallknecht et al. 1996, Haigh et al. 2002, Flacke et al. 2004, Sleeman et al.

2009). Both pathogens primarily affect white-tailed deer in North America, often leading to great morbidity and mortality (Stallknecht et al. 1996, Nol et al. 2009, Stallknecht et al. 2015, Stevens et al. 2015). Hemorrhagic disease is endemic to large regions of the

United States with multiple areas of enzootic stability (Stallknecht et al. 1996, Flacke et al. 2004). Of these two pathogens, BTV has been the most well studied due to the

17 detrimental effects of this pathogen on sheep, an important domestic ruminant primarily in Europe (Saegerman et al. 2008, Carpenter et al. 2009). While EHDV has not received the same level of attention historically, increased investigations into the ecology and epizootiology of this pathogen are ongoing as a result of increased economic impact of white-tailed deer farming (Anderson et al. 2017).

Deer farming is a growing industry in the United States with a total economic impact of up to $7.9 billion supporting greater than 56,000 jobs as of 2017 (Anderson et al. 2017). These values have grown considerably since 2007 when total economic impact was estimated at $3.0 billion supporting 29,000 jobs (Anderson et al. 2007).

EHDV negatively affects deer farms, with outbreaks occasionally leading to high mortality which discourages the movement of potentially susceptible deer into endemic areas (Haigh et al. 2002, Lee and English 2011). Due to the negative impact EHDV has on deer, both wild and farmed, there is increased pressure to investigate aspects of viral transmission, epizootiology, and ecology of the vectors.

Epizootic hemorrhagic disease virus is a double stranded RNA virus composed of a 10-segment genome (Huismans et al. 1979). Each segment encodes a separate protein including the structural proteins VP1-VP7 and the non-structural proteins NS1-

NS3 (Mecham and Dean 1988). The outer capsid is composed of proteins VP2 and

VP5, which bear serotype-specific antigens important for viral attachment to and infection of host cells (Mecham and Dean 1988, Mertens et al. 1989, Cheney et al.

1996). The inner capsid core is primarily composed of proteins VP3 and VP7 (Grimes et al. 1998). The remaining structural proteins, VP1, VP4, and VP6 act as viral transcriptase complexes (Anthony et al. 2009a). Non-structural proteins facilitate virus

18 movement and propagation within the cell. The NS1 protein is involved in the production of tubules to facilitate movement of the virus within the cell (Huismans and Els 1979,

Nel and Huismans 1991). NS2 is involved with the assembly of virus particles as an important part of the viral inclusion body (Theron et al 1996a; 1996b). Finally, NS3 and the NS3a subunit assist the virus in release from the host cell (Hyatt et al. 1991; 1993).

Currently, the only confirmed vector of EHDV in North America is C. sonorensis.

This species has been confirmed as a vector of all three EHDV serotypes currently in

North America, including EHDV-1, EHDV-2, and EHDV-6 (Foster et al. 1977, Jones et al. 1977, Ruder et al. 2016). While the range of C. sonorensis extends throughout the

United States, populations are sporadic east of the Mississippi river (Borkent and

Grogan 2009, Vigil et al. 2014) and multiple large-scale collection efforts in the southeastern United States have recovered few C. sonorensis individuals (Smith and

Stallknecht 1996, Smith et al. 1996b, Sloyer et al. 2018). Despite this, EHDV occurs in southeastern states, indicating that alternative vectors are likely transmitting EHDV in this region (Prestwood et al. 1974, Stallknecht et al. 1991, Murphy et al. 2006).

Various alternative vector species have been proposed, largely based on their great abundance and a close association with EHDV host species. These include species such as Culicoides debilipalpis Lutz, Culicoides obsoletus Meigen, Culicoides paraensis Goeldi, Culicoides spinosus Root and Hoffman, and Culicoides stellifer

(Coquillett) (Mullen et al. 1985a, Smith and Stallknecht 1996, Smith et al. 1996b). While abundance and association with viral hosts is an important aspect of implicating a vector species, these associations are not definitive evidence for vector incrimination.

Standardized criteria have been developed for the implication of arboviral vectors.

One such standard is the World Health Organization vector incrimination criteria.

These criteria include 1) recovery of virus from wild-caught specimens free from visible blood, 2) demonstration of the ability to become infected by feeding on a viremic vertebrate host or an artificial substitute, 3) demonstration of the ability to transmit the virus biologically by bite, and 4) field evidence confirming significant association of the infected arthropods with the appropriate vertebrate population in which disease is occurring (World Health Organization 1967). While criteria one and four rely on field collected evidence, criteria two and three are aspects of traditional vector competence experiments. These experiments are challenging with Culicoides as few species have been colonized and coercing uncolonized individuals to blood feed in a laboratory is challenging (Mullen et al. 1985b, Venter et al. 2005).

Very few vector competence experiments have successfully been carried out with Culicoides species present in the southeastern United States. Culicoides debilipalpis (as Culicoides lahillei) has been evaluated for vector competence to EHDV-

2 and was found to become infected when provided a high viral titer (Log10 5.3-6.0

TCID50 (Median Tissue Culture Infectious Dose)) but had low competence when provided low viral titer blood meals (Log10 2.1-3.0 TCID50) (Smith et al. 1996a).

Culicoides debilipalpis and C. stellifer also have been tested for vector competence for

BTV in one previous study, both through membrane feeding and intrathoracic inoculation. The study found low infection rates for both species, although one C. stellifer pool and one C. debilipalpis female inoculated intrathoracically as well as one viral blood-fed C. debilipalpis were positive indicating low level competence for infection

(Mullen et al. 1985b). One New York population of Culicoides venustus has been tested

20 for competence to EHDV-1 and EHDV-2 as well as to multiple strains of bluetongue virus. Very little evidence of infection was found for all strains assayed (Jones et al.

1983). For each of these competence assays, transmission potential was not investigated, which prevents the full calculation of vector competence for all these species.

While an important indication of a species ability to become infected with and transmit a virus, vector competence is only one aspect of vectorial capacity (C), which is an equation showing the efficiency of transmission of a vector-borne pathogen through calculating the number of infectious arthropods arising from a single infected host.

Originally developed for use with Anopheles mosquitoes that transmit malaria

(MacDonald 1957), this equation (1-1) takes into consideration vector competence (b), the extrinsic incubation period (n), biting rate of the vector (a), the probability of daily survival of the vector (p), and the vector density (m) (Kramer and Ciota 2015).

푚푎2푝푛푏 C= (1-1) −ln(푝)

This equation takes into consideration not only the intrinsic factors affecting the vector’s ability to transmit a virus, but also the biology and ecology of a vector population. This equation is an important foundation for the link between virus epidemiology, vector biology, and vector ecology.

One important aspect of Culicoides ecology that warrants additional investigation is that of the host use and blood-feeding preferences of Culicoides in North America.

Historically, most of the information gathered on host use was through the use of baited drop traps and aspiration of hosts (Humphreys and Turner 1973, Koch and Axtell 1979,

Schmidtmann et al. 1980, Raich et al. 1997). These studies provided a wealth of information prior to the availability of genetic methods; however, the limited host range that could be inferred from these studies made their applications somewhat limited. With the advent of PCR-based blood meal analysis, blood-meal identification can allow researchers to identify novel reservoir hosts (Kirstein and Gray 1996, Haouas et al.

2007), potential arboviral vectors (Molaei and Andreadis 2006, Molaei et al. 2009), and better understand transmission dynamics of pathogens (Molaei et al. 2010, Estep et al.

2011). Host use patterns of Culicoides could, therefore, be used to identify potential vector species by investigating those species that feed preferentially on EHDV hosts.

Daily survivor of the vector (p) is also an important component of the vectorial capacity equation as it is considered in both the numerator and denominator of the equation. The average adult lifespan of Culicoides is challenging to estimate. A study of

C. sonorensis found that larger flies, produced by rearing midges in lower density conditions, resulted in longer-lived individuals than those reared in high density conditions (Akey et al. 1978). Temperature also has a significant effect on the longevity of adult Culicoides, with longevity decreasing with increasing temperature (Wellby et al.

1996, Wittman et al. 2002, Lysyk and Danyk 2007). Average adult longevity in field populations is extremely challenging to estimate due to variability in these factors.

The lifespan of a midge also is dependent upon their ability to attain adequate energy stores (carbohydrates). In the laboratory, Culicoides nebeculosus females fed on sugar and blood survived on average 18.2 days longer and produced 20.1 more eggs than cohorts fed on blood alone (Kaufmann et al. 2015). Male and female

Culicoides have been found approaching and feeding on flower nectar (Downes 1958).

Additionally, multiple studies have used anthrone to test for the presence of fructose and sucrose in collected individuals. One study found that over 60% of both Culicoides melleus (Coquillett) and Culicoides hollensis (Melander and Brues) were positive for sugar indicating frequent sugar feeding in nature (Magnarelli 1981), while other studies found sugar feeding rates of up to 80% within the Obsoletus and Pulicaris groups

(Kaufmann et al. 2015). Greater sugar feeding rates have been found in parous females than nulliparous females and males indicating the greater role sugar plays during this stage (Mullens 1985, Stewart and Kline 1999). There is evidence that some nulliparous

Culicoides females do not need to sugar feed, relying solely on larval energy reserves and an initial blood meal to produce the first egg batch; however, for the second gonotrophic cycle, sugar-feeding becomes necessary to accumulate adequate resources for egg production (Mullens and Schmidtmann 1982, Stewart and Kline

1999).

In most environments, all necessary resources for Culicoides survival and reproduction are not available in a single area requiring midges to travel between breeding sites, hosts, oviposition sites, sugar sources, and resting locations. Mark- release-recapture (MRR) studies have provided some insight into the horizontal flight distances of Culicoides. One study on Palearctic Culicoides species identified travel less than 1.00km in males and 2.21km on average for females (Kluiters et al. 2015). A study conducted on C. sonorensis (as Culicoides variipennis sonorensis) have identified maximum travel distances of 0.80km in males and 4.00km in females, although average travel for all recovered individuals was only 1.89km from the release site (Lillie et al.

1981). Similarly, mean distance traveled by C. mississippiensis was 2.00-2.20km (Lillie

23 et al. 1985). Longer-distance travel on air currents is believed to occur, although this is thought to be entirely incidental (Ducheyne et al. 2007, Sanders et al. 2011).

In addition to horizontal movement on the landscape, Culicoides have been found to move vertically within the forest canopy. One study in the Amazon found

Culicoides species stratifying vertically in traps placed at 1m, 5m, and 10m, with the greatest diversity collected at 5m (Veras and Castellon 1998). Vertical stratification has been investigated in some North American species with species such as C. furens, C. haematopotus, and C. stellifer occurring in greater abundance in the forest canopy than below, while species such as C. obsoletus and C. travisi were more abundant below the canopy than above (Henry and Adkins 1975). The impetus for these transitions between ground and canopy are unclear and may be linked to host preferences (Swanson et al.

2012), searching for sugar sources (Ulyshen 2011), or finding shelter (Service 1971).

Understanding the dispersal ecology of Culicoides is vital for optimizing sampling protocols that will generate accurate data on the abundance and survival of vectors.

There are many aspects of the ecology of Culicoides and the epizootiology of

EHDV that remain unknown. In order to develop sound management strategies for

EHDV on deer farms and big game preserves in Florida, the likely vectors of EHDV need to be identified. This will be achieved through the combined field and laboratory investigations of Culicoides ecology and EHDV infection with the end goal of identifying new vector species and candidates.

CHAPTER 2 HOST USE PATTERNS OF CULICOIDES SPP. BITING MIDGES AT A BIG GAME PRESERVE IN FLORIDA, USA1

Culicoides spp. biting midges (Diptera: Ceratopogonidae) damage livestock and wildlife through transmission of numerous viruses, particularly from the genus Orbivirus.

These viruses include African horse sickness virus (AHSV), bluetongue virus (BTV), and epizootic hemorrhagic disease virus (EHDV). These three viruses can cause morbidity, and often mortality, in economically valuable livestock species such as cattle, horses, sheep, and deer. Outbreaks of these viruses can result in significant economic losses due to animal deaths, causing financial stress in areas where the economy is driven by livestock industries (Barnard et al. 1998, Kedmi et al. 2010b, Rushton and

Lyons 2015).

Culicoides-transmitted orbiviruses that cause hemorrhagic disease (BTV and

EHDV) are an important source of mortality to white-tailed deer (Odocoileus virginianus)

(Nol et al. 2010, Stevens et al. 2015). This cervid constitutes a large part of the burgeoning big game industry in the United States (Anderson et al. 2007, Adams et al.

2016). At least 5,555 breeding facilities for captive white-tailed deer are present in the eastern United States, with more than 300 of these found in the state of Florida (Adams et al. 2016). As of 2007, the cervid farming industry contributed three billion USD towards the United States economy and that value has likely risen (Anderson et al.

2007). Many deer farming operations cater to hunters and offer opportunities for hunting white-tailed deer and other cervids, such as elk (Cervus canadensis), axis deer (Axis axis), fallow deer (Dama dama), and exotic bovids such as nilgai (Boselaphus

1 Reprinted with permission from John Wiley & Sons Publishing 25 tragocamelus), and blackbuck (Antilope cervicapra). Operations dedicated to breeding, meat production, and other products are also common in many states (Anderson et al.

2007). Due to the significant impact that hemorrhagic disease has on these animals and the industry, a better understanding of the ecological interactions between the numerous biting midge species and diverse hosts on properties where multiple game species are present is needed.

Host associations for many of the 150 Nearctic Culicoides spp. (Borkent &

Grogan 2009), including the 49 known species present in Florida (Grogan et al. 2010), are poorly understood. Much of the information regarding host usage by Culicoides spp. is based upon specimens collected from animal-baited traps and/or observational studies (Hair & Turner 1968). While these studies can provide valuable information on whether a certain Culicoides spp. will feed upon a particular host, they cannot infer host breadth, as only one or a few (usually domesticated or tame vertebrate animals) are available for experimentation. For example, one study found that Culicoides kibunensis

Tokunaga constituted the majority of trap captures in bird-baited traps, seemingly indicating ornithophily in this species (Synek et al. 2013). However, other studies utilizing polymerase chain reaction (PCR) based blood meal analysis found that C. kibunensis fed on humans (Santiago-Alarcon et al. 2012, Santiago-Alarcon et al. 2013) and cows (Lassen et al. 2012). Further, animal-baited traps can often provide conflicting results. Two separate animal-baited studies were used to investigate host preference for Culicoides sanguisuga (Coquillett). One study identified preference for large mammals (Humphreys and Turner 1973) while the other identified a preference for

26 poultry (Greiner et al. 1978). Results such as these confound our understanding of selection in this species.

Blood meal analysis of field-collected engorged females through PCR amplification of host DNA present in a blood meal is a valuable method to determine host use of many blood feeding arthropods, including biting midges (Slama et al. 2015), ticks (Allan et al. 2010), sandflies (Chaskopoulou et al. 2016), tsetse flies (Muturi et al.

2011), and mosquitoes (Burkett-Cadena et al. 2008) among others. Blood meal analysis targets regions of certain genes that are well conserved throughout the animal kingdom but show sufficient interspecific variation for identification at the species level, including

16s rRNA (Sarri et al. 2014), cytochrome b oxidase (Garros et al. 2011, Slama et al.

2015), and cytochrome c oxidase I (Ferri et al. 2009). Since contact (biting) between vectors and susceptible hosts is a critical variable in determining vectorial capacity of a putative vector species, clearly delineating patterns of host use is important for inferring vector status.

Forage ratios use data on relative abundance and host use of various vertebrate species to infer the propensity of blood-feeding arthropods to feed on specific animal species (Hess et al. 1968). This basic metric has been used extensively to investigate patterns of host use for vectors of diverse pathogen systems, including malaria (Parida et al. 2006, Lardeux et al. 2007), leishmaniasis (Agrela et al. 2002, Rossi et al. 2007), and numerous arboviruses (Hess & Hayes 1970; Braverman et al. 1971; Ponlawat &

Harrington 2005; Samuel et al. 2008). Forage ratios provide a useful general approximation of whether host species are utilized at a rate different than their relative abundance within a community (Hess et al. 1968). Forage ratios are typically calculated

27 at a landscape scale, such that natural movement and aggregation of both vectors and hosts shape the outcome of the interaction, i.e., host use (Chaves et al. 2010). Host species that are utilized disproportionately greater than their relative abundance are considered to be “preferred” by the blood-feeder, while those utilized disproportionately lesser than their relative abundance are considered to be “avoided”. Host preference and avoidance do not quantify small-scale interactions of blood-feeders and their hosts, such as behavioral avoidance or defensive behaviors of host animals, but instead, simply quantify the outcomes of these multi-species interactions on a larger community- level scale.

No data are currently available on Culicoides host preference of exotic ungulates such as fallow deer, axis deer, elk, and Père David's deer that are often intermixed with native species, such as white-tailed deer, on hunting preserves (Anderson et al. 2007).

Because ungulate density on these hunting preserves often far surpasses bovid or cervid densities in more natural environments, the increased host abundance could result in increased Culicoides spp. density (Garci-Saenz et al. 2011). Additionally, the availability of exotic host species as resources for blood meals or sources of arboviruses could change Culicoides spp. composition, host use, or virus transmission dynamics. If ungulate species composition ultimately plays a role in Culicoides spp. ecology, and even affects hemorrhagic disease transmission, then it may be possible to reduce transmission by changing host community composition.

The goals of this research were to (1) quantify host use and host preference of

Culicoides spp. on a big game preserve in the Florida panhandle; and (2) use

Culicoides spp. abundance and host preference results to draw conclusions regarding

28 candidate vectors of hemorrhagic disease in Florida. Culicoides sonorensis Wirth &

Jones, the only confirmed vector for EHDV in North America, is considered rare in

Florida. Due to the persistent transmission of these arboviruses in the region (Ruder et al. 2015a), it is likely that other vector species are active in the area. These data can be used by land managers and researchers to focus control efforts towards biting midge species that are most pestiferous towards captive animals and may also be candidate vectors for hemorrhagic disease in Florida.

Methods

Field Methods

The site for this study was a privately owned 200-hectare big game hunting preserve and white-tailed deer breeding farm located in Gadsden County, Florida, USA.

A variety of habitat types were present on the property including hardwood forest, hardwood swamp, mixed pine forest, open fields, and large ponds (Figure 2-1). A diverse assemblage of big game species (Bovidae and Cervidae) were present. Bovids on the property included blackbuck antelope (Antilope cervicapra), nilgai (Boselaphus tragocamelus), goats (Capra hircus), waterbuck (Kobus ellipsiprymnus), scimitar-horned oryx (Oryx dammah), gemsbok (Oryx gazella), and bighorn sheep (Ovis aries).

Cervidae on the property included axis deer (Axis axis), North American elk (Cervus canadensis), sika deer (Cervus nippon), sika deer hybrids (Cervus nippon x Cervus canadensis), Père David's deer (Elaphurus davidianus), fallow deer (Dama dama), and white-tailed deer (Odocoileus virginianus) (Table 2-1). The property consisted of one large open preserve on which most of the animals were free to roam (Figure 2-1, orange lines indicate property perimeter), but also contained two penned areas where

29 approximately half of the property’s white-tailed deer were kept (Figure 2-1, red lines indicate white-tailed deer pens).

Twenty Centers for Disease Control and Prevention (CDC) miniature light traps

(Model 2836BQ, BioQuip, Rancho Dominguez, CA, USA) with black light LED arrays

(Model 2790V390, BioQuip, Rancho Dominguez, CA, USA) were hung from 1.63m tall shepherd's hooks at 1.37m in height. Trap locations were selected using the random point generator in ArcGIS v 10.3 (ESRI, Redlands, CA). Sets of 20 random points were iteratively generated and tested for spatial randomness using the average nearest neighbor index in ArcGIS. The final set of trap sites was spatially random and represented all habitat types in the study area (Figure 2-1). Once selected, trap site locations did not change throughout the study. Trap chambers were modified with mesh

(3 x 3mm) to exclude large arthropods, and a cloth funnel was used to direct small arthropods into a 50mL conical tube at the bottom of the trap containing 90% ethanol.

Disturbance of the trap wiring by animals on the preserve was avoided by wrapping heavy-duty garden hose around the trap wires. Traps were powered via 6V-12Ah gel- sealed battery controlled by a timer to operate between one hour prior to sunset and one hour after sunrise. Total trap run time varied between 12 hours during the summer to 15 hours during the winter.

Trapping was conducted twice weekly between July 2015-December 2016.

During the months of November-March, the number of traps operated on the property decreased to 10. The ten traps that remained operational during the over-wintering period were selected to ensure sampling at all major habitats on the property. These traps were furthermore selected because they yielded the greatest Culicoides spp.

30 abundance and diversity during the previous season (summer of 2015). Protocols for operation of these ten traps were not changed.

Collections were also made by aspiration from bottle-raised, tame adult white- tailed does from July 2015 through September 2016 from the pens. Aspirations were conducted for 10 minutes at three time periods (dawn, midday, and dusk) once per week on any approachable deer in the pen. The aspirator design was an acrylic tube

(8.89cm diameter) containing a computer cooling fan powered by a 12V battery. Midges were collected into removable collection cups, which were stored at -20˚C until identification.

Laboratory Methods

All Culicoides spp. specimens collected were identified to species using morphological identification keys in Blanton and Wirth (1979). Blood-engorged females were each placed into individual 1.5mL DNase/RNase-free microcentrifuge tubes and stored at -20˚C. Total DNA was extracted from individual biting midges using Chelex resin or Instagene (BioRad Inc., Hercules, CA). DNA Extraction using Chelex followed protocols outlined in Fabian et al. (2004). In brief, a sterile pestle was used to homogenize the sample in 10µL of 0.9% NaCl, followed by addition of 5% Chelex suspension (240µL at 100°C), incubation at 100°C for 10 min, then centrifugation for 5 min at 3,099g. Supernatant was collected into a sterile 1.5mL tube. Instagene protocols were identical, except pre-warming of Instagene to 100°C was not required.

Regions of extracted DNA were amplified with PCR using three primer sets

(Blosser et al. 2016) targeting different vertebrate taxa. All samples were run initially on the mammalian/amphibian primer set (F: CTCCATAGGGTCTTCTCGTCTT, R:

GCCTGTTTACCAAAAACATCAC). If no amplification was observed, samples were subsequently run on the reptile (F: CTGACCGTGCAAAGGTAGCGTAATCACT, R:

CTCCGGTCTGAACTCAGATCACGTAGG) and avian (F:

GGACAAATATCATTCTGAGG, R: GGGTGGAATGGGATTTTGTC) primer sets.

Reagents used per sample for all PCR protocols included 14.25µL molecular biology grade water, 2.5µL 200mM Tris-HCl reaction buffer, 2.5µL 2mM dNTPs, 1.5µL 50mM

MgCl2, 0.625µL 20µM forward primer, 0.625µL 20µM reverse primer, and 0.5µL Taq polymerase (Invitrogen, Thermo Fisher Scientific Inc., Waltham, MA). Addition of 2.5µL of extracted DNA (50-150ng/µL concentration) resulted in a total reaction volume of

25µL per well. Cycling conditions for the mammalian/amphibian assay were 94°C for 2 min, followed by 35 cycles of 94°C for 30 s, 57°C for 30 s, and 72°C for 1 min. Cycling conditions for the reptile specific primers were 94°C for 2 min followed by 35 cycles of

94°C for 30 s, 62.5°C for 30 s, and 72°C for 1 min. The cycling conditions for the avian primers were 94°C for 2 min, followed by 35 cycles of 94°C for 30 s, 60°C for 30 s, and

72°C for 1 min. All PCR runs included a negative control (molecular grade water) to monitor for contamination.

PCR products were visualized on 1% agarose gels with electrophoresis at 100V for 35 min. A blue light transilluminator was used to visualize bands and all positive amplicons of the appropriate size were directly sequenced at a commercial laboratory using Sanger sequencing (Eurofins Genomic, Louisville, KY). Resulting sequences were compared with sequence information in GenBank (National Institutes of Health:

National Center for Biotechnology Information) using BLAST (Basic Local Alignment

Search Tool), and samples with ≥95% identity matches and ≥75% query coverages

32 were retained. Samples that failed to provide high identity matches were re-run and the resulting PCR product was purified (NucleoSpin Gel and PCR Clean-up kit, Macherey-

Nagel, Bethlehem, PA) prior to resubmission for sequencing.

Data Analysis

Forage ratios (FR; relative host use/relative host abundance) were calculated to determine preference for various ungulate species and families on the property (Hess et al. 1968, Manly et al. 2007). Relative host use was determined for each Culicoides species by dividing the total blood meals taken by that Culicoides species from a particular host organism by the total number of blood meals taken by that Culicoides species overall. Relative host abundance was calculated by dividing the abundance of a specific host species on the property by total abundance of game animals present using information provided by the game preserve manager (Table 2-1). Host abundance data were available for the end of each year. For this reason, data for 2015 and 2016 were analyzed separately. FR within the 95% confidence intervals were considered to show preference or avoidance (equations 4.14 and 4.15 in Manly et al. 2007). FR greater than one indicated preference for a host species, while FR less than one indicated avoidance of a host species. Avoidance and preference values were not calculated for rare host species, i.e., less than five individuals, which included the sheep, waterbuck, and goats

(Lechowicz 1982). Because both Culicoides spp. and vertebrate hosts may differentially utilize available habitat types, we performed linear regression analysis to compare the host use between dominant habitat types. Habitat classes were derived from the Florida

Habitat and Land Cover dataset available through the Florida Geographic Data Library

(www.fgdl.org) and were developed by the Florida Fish and Wildlife Conservation

Commission using Landsat Enhanced Thematic Mapper satellite imagery. For this study, we simplified the 11 original classes into five classes based on field observation and ground truthing (Munga et al. 2006, Steiger et al. 2012). Separate analyses were conducted for C. debilipalpis and C. stellifer comparing proportion of bloodmeals from the various vertebrate hosts between the three dominant habitat types: hardwood bottomland, upland pine, and ecotone. This analysis allowed us to validate that calculated preserve-wide FR were representative of all three major habitat classes and were not biased by collections from a few dominant trap locations. Sample sizes for

Culicoides spp. other than Culicoides debilipalpis Lutz and Culicoides stellifer

(Coquillett) were too small to permit robust comparisons.

A chi-square goodness of fit test (R Core Team 2016, R 3.3.2 software, Vienna,

Austria) was used to investigate differences in the community composition of total

Culicoides spp. of all physiological statuses collected in traps, total blood meals taken from white-tailed deer, and of midges aspirated directly from white-tailed deer. The analysis was limited to the ten most commonly sampled Culicoides spp. and was performed in order to determine whether the different collection methods used provided comparable results for these species.

Results

Forage Ratios

During the full trapping period (July 2015-December 2016), a total of 69,964

Culicoides were collected, including 2,143 blood-engorged Culicoides. Out of the blood- engorged midges analyzed, 78.6% yielded sequences with ≥95% identity match to vertebrate hosts in GenBank. In 2015 and 2016, 933 of 1,201 (77.7%) and 753 of 942

(79.9%) blood meals were successfully identified, respectively. Our analysis was unable to differentiate species within the genus Cervus (i.e., elk, sika deer, and sika deer x elk hybrids). For this reason, these species were combined into a single group.

Although some Culicoides spp. fed on birds and a single amphibian blood meal was identified, the vast majority (99.4%) of blood meals were from mammalian hosts

(Table 2-2). Of the total mammalian blood meals 3.6% in 2015 and 5.0% in 2016 were derived from non-ungulate mammals (i.e., eastern gray squirrel Sciurus carolinensis, northern raccoon Procyon lotor, striped skunk Mephitis mephitis, domestic dog Canis lupus familiaris, and human Homo sapiens; Table 2-3), for which relative abundances were not quantified. For these reasons, forage ratio calculations were restricted to species of Cervidae and Bovidae identified in blood fed Culicoides spp.

The majority of successfully identified blood meals in both years were from C. stellifer, with 850 of the total blood meals (91%) in 2015 and 611 of the total blood meals (81%) in 2016 derived from this species (Table 2-4). Host species were variously preferred or avoided by C. stellifer and C. debilipalpis. While host species were preferred or avoided similarly between years in some instances, exceptions occurred where FR varied between preference and avoidance for the same game species between 2015 and 2016 (Figure 2-2). For example, C. debilipalpis showed a preference for Père David's deer in 2015, however this preference was not observed the following year. During both years, C. debilipalpis preferred white-tailed deer and avoided blackbuck, gemsbok, nilgai, scimitar-horned oryx, and fallow deer. All other hosts of C. debilipalpis were neither preferred nor avoided. Culicoides stellifer showed a preference for fallow deer in 2015; however, preference was not observed the following year. Both

35 years, C. stellifer showed a preference for Cervus species. In 2016, C. stellifer avoided gemsbok, but avoidance was not observed in 2015. Culicoides stellifer avoided blackbuck and axis deer during both years. Culicoides stellifer showed no preference or avoidance of the other host species in this analysis (Figure 2-2).

Species such as Culicoides biguttatus (Coquillett) and Culicoides pallidicornis

Kieffer are predominantly spring species (Blanton and Wirth 1979) and as such, were only collected during 2016 (sampling began in July in 2015). Forage ratios for C. biguttatus (n=72) were largely equivocal, except for a preference for Père David's deer

(FR=11.3, SE=3.26). Culicoides pallidicornis (n=16) preferred white-tailed deer

(FR=1.54, SE=0.26) and avoided fallow deer, Père David's deer, blackbuck, gemsbok, and scimitar-horned oryx (FR=0 for each).

Forage ratios calculated to determine whether abundant Culicoides spp. showed preference for a specific mammalian family, Bovidae or Cervidae, indicated broad preference for Cervidae and avoidance of Bovidae in C. debilipalpis. Although C. stellifer also fed more heavily on Cervidae than Bovidae, 95% confidence intervals revealed their preference for Cervidae was not significant (Figure 2-2).

Habitat Associated Host Use

Host use was consistent between major habitat types for C. debilipalpis and C. stellifer (Table 2-5). Linear regression analysis demonstrated significant positive relationships between habitat types (hardwood bottomland, upland pine, ecotone) for the proportion of bloodmeals from various host species (Table 2-5). Host use from the two dominant yet qualitatively different habitats, hardwood bottomland and upland pine

36 habitats, showed significant positive relationships for both C. debilipalpis (R2=0.97, df=15, P<0.001) and C. stellifer (R2=0.77, df=15, P<0.001).

White-tailed Deer Aspiration

In all, 685 biting midges were aspirated from white-tailed deer. In a comparison of species composition and abundance for the ten most common Culicoides spp. of a) all Culicoides collected in traps, b) those species that fed only on white-tailed deer, and c) those that were aspirated off white-tailed deer, a few major differences were observed (Figure 2-3). While C. stellifer was the most abundant species present in total trap counts (n=47,180 individuals) and took the most blood meals from white-tailed deer

(n=555 blood meals), this species was not the most commonly collected species in aspirations (n=213 individuals, 31.1% of total collections). Instead, C. pallidicornis was the most abundant species caught in aspirations (n=373 individuals, 54.5% of total collections). Additionally, although C. haematopotus was the second most abundant species in total trap counts (N=9,709 individuals), only 13 blood meals were successfully analyzed from this species. Of these 13 blood meals, two (15.4%) originated from white-tailed deer, two (15.4%) from humans, one (7.7%) from Cervus spp., and six (46.2%) were from various birds (Table 2-3). Further, none were aspirated from white-tailed deer during this study. A chi-squared goodness of fit test indicated that the comparative abundance of the ten most common Culicoides species collected in total trap counts, identified through blood meal analysis feeding on white-tailed deer, and aspirated from white-tailed deer were significantly different (χ2=211.89, df =18,

P <0.001).

Discussion

This is the first in-depth Culicoides spp. blood meal analysis study conducted in

North America on a big game hunting preserve that houses a diversity of bovids and cervids and experiences cases of EHDV annually. A total of 1,686 blood meals were successfully analyzed allowing the calculation of forage ratios that can guide our understanding of Culicoides spp. ecology and host preference on this and similar properties in the southeastern United States.

Through the calculation of forage ratios, this work also allowed us to identify

Culicoides spp. that feed preferentially on hosts affected by BTV and EHDV, an important step in implicating candidate vector species. Species such as C. stellifer, C. debilipalpis, and C. pallidicornis fed on white-tailed deer and were aspirated from tame white-tailed deer more than other Culicoides spp. on the property (Figure 2-3). They were also some of the most abundant species in total trap counts. Culicoides pallidicornis was the most abundant early spring species, and C. stellifer was the most abundant summer and fall species. For these reasons, we have identified C. stellifer, C. debilipalpis, and C. pallidicornis as candidate vector species that warrant further investigation as EHDV vectors in northern Florida. Although C. haematopotus was the second most abundant midge species in total trap collections, few blood-engorged individuals were captured. Just under half of the blood meals from this midge species were taken from birds (6/13) while white-tailed deer constituted only 2/13 total bloodmeals. Aspiration collections reinforced these findings as no C. haematopotus were documented approaching white-tailed deer throughout the study period. These data

38 suggest that C. haematopotus is not a putative vector species for these orbiviruses in this region.

The vast majority of blood meals analyzed for this study originated from big game species at large in the study preserve. However, while half of Culicoides spp. in this analysis (7/14 species) predominantly took blood meals from preserve game species, several midge species appeared to avoid big game species in favor of non-game mammals and birds (Table 2-3). This suggests that species such as C. baueri, C. bickleyi, C. crepuscularis, C. haematopotus, and C. paraensis are not likely involved in transmission of EHDV and BTV on this property. However, previous studies have indicated that pooled specimens of C. crepuscularis and C. haematopotus have tested positive for BTV in Louisiana (Becker et al. 2010), so these data cannot rule out roles for these species in transmission of BTV. Additionally, despite being a confirmed BTV vector (Tanya et al. 1992), the low abundance of C. insignis on the property and lack of blood meals from preserve species indicates that C. insignis is also not a likely primary vector of BTV or EHDV on this property. However, it is also important to note that vectorial capacity can be affected by a variety of other factors, including the vector competence of the putative vector species, probability of the vectors surviving each day, and the extrinsic incubation period of the pathogen in the vector. For this reason, it is not possible to completely dismiss a potential vector species on the basis of abundance or biting rate on an affected host alone.

Spring species (C. biguttatus and C. pallidicornis) were only collected during one study year. For this reason, it is difficult to predict whether preferences shown by these two species would remain the same between years. However, the preference for white-

39 tailed deer by C. pallidicornis observed in this analysis could have implications for the maintenance of EHDV and BTV through winter and spring months, if this species were determined to be a competent vector. Although EHDV and BTV outbreaks generally occur in the late summer and fall (Ruder et al. 2015a), virus circulation has been observed on this property over winter and into spring. Our findings provide a first glimpse at possible maintenance mechanisms that allow these viral agents to persist on the landscape through the winter. Low titer viremia for EHDV can persist for up to 59 days post inoculation in white-tailed deer (Ruder et al. 2015a). Therefore, a deer infected in late fall or early winter could maintain viremia through the winter in Florida until emergence of early spring species such as C. pallidicornis that then transmit the virus to other hosts.

A strong preference for cervids and avoidance of bovids by C. debilipalpis was observed during both years. These preference patterns, as well as all the forage ratios discussed, should be interpreted carefully as they could be affected by a variety of exogenous forces such as olfactory cues, spatial distribution, animal abundance fluctuations, or defensive behaviors. Olfactory cues between cervids, bovids, or the individual species themselves could differ and, therefore, lead to diverse responses by different Culicoides spp. (Isberg et al. 2013). Skin glands that produce volatile compounds are common in many cervid (Gassett et al. 1996; 1997, Wood 2003,) and bovid (Wood 1997;1998) species for intraspecific communication (Muller-Schwarze et al. 1984). The specific effect of many of these compounds on Culicoides spp. behavior is unknown. For this study, Y-tube olfactometer analysis was not conducted to look for olfactory preferences, which leaves uncertainty about the true source of the preference

40 patterns seen. In addition to olfactory cues, the spatial distribution and resource selection by host animals, and spatial arrangement of midges, on the property likely led to the preference for cervids over bovids, as well as to observed preferences for specific species. One example of this phenomenon observed during this study is the preference of C. biguttatus for Père David's deer. This deer species is often associated with aquatic habitats (Li et al. 2007), and all the Père David's deer blood meals taken by C. biguttatus were from traps located near (<3m away from) a large stream. The resulting forage ratio is likely due to the overlapping spatial distribution and similar habitat preferences of this midge and deer species rather than an innate host preference.

Forage ratios also could be affected by fluctuating animal abundances between study years and habitat associations of various host species, as well as the species of

Culicoides. Births and deaths of free-ranging animals on the property, as well as their seasonal and circadian movements between habitats will certainly affect the accuracy of forage ratios. However, our analysis comparing host use between habitat types (Table

2-5) indicates that the patterns of host use exhibited by the most abundant species, C. debilipalpis and C. stellifer, are robust to habitat type. This represents an important step in understanding the role that environment plays in host preference, which is an important component of transmission cycles of vector-borne disease. Finally, defensive behaviors by host animals are known to alter host use by blood feeding arthropods and may have led to the observed patterns. The use of behaviors such as foot stomps, wing shakes, and head movements by birds (Darbro and Harrington 2007) and ear twitches, hindfoot scratching, and head and body shakes by small mammals (Walker and Edman

1986) has been shown to deter mosquitoes, often causing them to move to host species

41 that use fewer of these defenses. Studies on mosquitoes have found that regardless of innate preference, mosquitoes are ultimately most likely to take their blood meal from a tolerant host from which it experiences the least defensive behavior (Edman et al.

1974). Different host defensive behaviors could be utilized by bovids and cervids that result in the comparably heavy use of cervids in this analysis. Although both groups use behaviors such as ear flicking, longer tails in large bovids have been proposed as an evolved mechanism for defending against biting flies (Mooring et al. 2007). The tails of cervids are typically shorter than many bovid species and, therefore, may not provide as much protection from insect pests. The authors acknowledge the many possible reasons for the patterns seen in this analysis and feel that additional studies are necessary to quantify forces that may drive observed patterns of preference and avoidance. Interpretation of avoidance is particularly challenging, given the lower likelihood of less-abundant animals to aggregate in areas near traps, perhaps causing their blood meals to be underrepresented using our methods.

Although our findings indicated that C. stellifer did not prefer white-tailed deer relative to their abundance on the property, this may not be justification for dismissing the potential for this species to serve as a vector of EHDV. Based on total trap collections, C. stellifer was the most abundant species present on the property and fed heavily on white-tailed deer throughout the study period. Despite avoiding white-tailed deer relative to their abundance, they do feed heavily on this host species and would, therefore, be a likely candidate vector species. In addition, other large ruminant mammals for which C. stellifer does show a preference may also play a part in the transmission cycle of these viruses. Infection studies indicate that elk (Hoff and Trainer

1973) and fallow deer (Gibbs and Lawman 1977), both of which were present on this preserve, can become infected and viremic with EHDV without exhibiting clinical signs

(Gibbs and Lawman 1977). Our data suggest that C. stellifer could be transmitting virus between viremic exotic host species and white-tailed deer.

Our study revealed some important limitations of different trapping methods. We observed differences in blood fed Culicoides species composition and abundance in light traps versus Culicoides spp. collected via aspiration. Aspiration from tame deer revealed the limitations of trap data to inform vector status, specifically in relation to C. pallidicornis. Although this species made up a relatively small percentage of the total midges from light traps and white-tailed deer blood meals on the property, C. pallidicornis constituted the majority of midges taken from aspirations. This informs us that this species readily approaches and feeds on white-tailed deer but does not often enter traps after blood feeding. For this reason, forage ratios for C. pallidicornis may have been underestimated. A chi-squared goodness of fit test indicated that there were significant differences in capture rates among trap types for the 10 Culicoides spp. analyzed. This result emphasizes the importance of using multiple methods of collection when implicating putative vector species. It should be noted, however, that aspirations were only carried out in two pens on the property (constituting less than 10% of the property size) and were, therefore, not spatially arranged in a manner similar to the light traps. This limited sampling could have resulted in biased sampling of the species willing to approach the tame deer in the pen environment versus other habitats on the property. An additional limitation of this study was undocumented wildlife populations around the property perimeter that might have biased parametrization of available

43 hosts. The area surrounding this preserve was mostly rural forestland with interspersed homes. The property shared a short border with a state forest. Wild deer in the vicinity of this property could have led to lower calculated forage ratios for white-tailed deer than we identified in this study if midges were feeding on wild deer located beyond the fence line. We did not conduct surveys into non-game species abundance on or off the property. Relatively few blood meals (n=80) were taken from non-game mammal and bird species, so we were unable to draw any significant conclusions from these data.

In conclusion, the calculation of forage ratios for common Culicoides spp. permitted better documentation of the ecology of some of the most abundant Culicoides spp. in the Florida panhandle. These forage ratios, along with information from total midge abundance in traps and tame deer aspirations, allowed us to draw inferences about potential vector species on the property. The abundance of C. stellifer and their numerous interactions with white-tailed deer and other EHDV-susceptible hosts makes them an important candidate vector species, warranting vector competence studies.

Similarly, the preference of C. debilipalpis for white-tailed deer as well as their presence in aspirations from tame deer indicates that this is also a species of interest that should be investigated. Finally, C. pallidicornis is a highly abundant early spring species that preferred white-tailed deer and was caught in great abundance during aspirations.

Although their emergence did not coincide with peak seasonal outbreaks of BTV and

EHDV, C. pallidicornis abundance during the spring season could provide a possible overwintering mechanism for hemorrhagic viruses and warrants investigation.

Figure 2-1. Map of the big game preserve located in Gadsden County, FL and Culicoides spp. by habitat type. The perimeter of the property is indicated in orange, the perimeter of the two penned areas containing white-tailed deer is indicated in red and paved roads in black. Pie charts represent the percentages of blood-engorged midges captured by traps at that point. Pie charts with an asterisk indicate the 10 traps that were run continuously throughout the study. The other 10 traps were not run between November- March. Aside from the two penned areas of white-tailed deer kept for breeding purposes, animals (including additional white-tailed deer) were free to roam throughout the property within the property limits.

Table 2-1. Total big game abundance estimates on the Gadsden County, Florida big game preserve during the 2015 and 2016 trapping season. Abundance estimates were determined in December of each year by the property manager. Family Scientific Name Common Name 2015 2016 Abundance Abundance Bovidae Antilope cervicapra Blackbuck 40 30 Bovidae Boselaphus tragocamelus Nilgai 6 8 Bovidae Capra hircus Goat 3 3 Bovidae Kobus ellipsiprymnus Waterbuck 1 1 Bovidae Oryx dammah Scimitar-horned Oryx 6 8 Bovidae Oryx gazelle Gemsbok 7 9 Bovidae Ovis aries Bighorn Sheep 3 3 Cervidae Axis axis Axis Deer 40 40 Cervidae Cervus spp. Elk/Sika/hybrids 19 22 Cervidae Dama dama Fallow Deer 12 24 Cervidae Elaphurus davidianus Père David’s Deer 7 9 Cervidae Odocoileus virginianus White-tailed Deer 130 148 Total 274 305

Table 2-2. Total bloodmeals taken by all Culicoides spp. on game mammals, non-game mammals, birds, and other sources on the study preserve in 2015 (July- December) and 2016 (January-December). Year Game Non-game Birds Other Total Mammals Mammals 2015 894 33 5 1 933 2016 708 38 4 3 753 Total 1602 71 9 4 1686

Figure 2-2. Forage ratios (±95% confidence intervals) for ungulate species and family by Culicoides debilipalpis (upper panel) and Culicoides stellifer (lower panel) for 2015 (in red) and 2016 (in blue) from a big game preserve in Gadsden County, FL.

Table 2-3. Non-game mammalian and avian blood meals taken by Culicoides spp. during the full study period (July 2015-December 2016). Culicoides were collected at a big game preserve in Gadsden County, FL using Centers for Disease Control and Prevention (CDC) miniature light traps.

Cow Dog Human SquirrelGray Raccoon Skunk Striped Eastern Crow American Chicken Yellowthroat Common Mississippi Kite Cardinal Northern Red Vulture Turkey

eyed Vireo

C. arboricola 0 0 0 0 0 0 0 1 0 0 0 0 1 (n=5) C. baueri 0 0 0 1 0 0 0 0 0 0 0 0 0 (n=1) C. biguttatus 0 0 2 0 0 0 1 0 0 0 0 0 0 (n=75) C. crepuscularis 0 0 1 1 0 0 0 0 0 0 0 0 0 (n=2) C. debilipalpis 0 0 2 1 0 0 0 0 0 0 0 0 0 (n=70) C. haematopotus 0 0 4 0 0 0 0 0 1 1 3 1 0 (n=13) C. hinmani 0 0 0 1 0 0 0 0 0 0 0 0 0 (n=2) C. insignis 0 0 1 0 0 0 0 0 0 0 0 0 0 (n=1) C. pallidicornis 0 0 1 0 0 0 0 0 0 0 0 0 0 (n=17) C. paraensis 0 0 0 2 1 0 0 0 0 0 0 0 0 (n=3) C. stellifer 7 1 42 1 1 1 0 0 0 0 0 0 0 (n=1,461) Total 7 1 53 7 2 1 1 1 1 1 3 1 1

Table 2-4. Total game mammal (Cervidae and Bovidae) blood meals taken by Culicoides spp. for which >5 game mammal blood meals were recovered. Culicoides were collected using Centers for Disease Control and Prevention (CDC) miniature light traps on a big game preserve in Gadsden County, Florida in 2015 (July-December) and 2016 (January-December). 2015 2016 Family Blood Meal C. debilipalpis C. stellifer C. venustus C. biguttatus C. debilipalpis C. pallidicornis C. stellifer C. venustus Host

Cervidae Axis deer 2 14 0 6 0 1 36 1

Elk/Sika deer 6 276 1 14 4 1 141 2

Fallow deer 0 206 1 9 0 0 56 2

Père David’s 3 17 1 24 0 0 43 0 deer White-tailed 42 276 9 16 10 12 279 9 deer Bovidae Blackbuck 0 7 1 1 0 0 7 2

Gemsbok 0 0 0 0 0 0 4 1

Nilgai 0 10 0 1 0 2 9 1

Scimitar- 0 19 1 1 0 0 5 0 horned oryx Annual Total 53 825 14 72 14 16 580 18

Table 2-5. Results of linear regression testing for differences in host use (proportion of blood meals from each host species) by Culicoides spp. among major habitat types. Species Comparison R2 df P C. debilipalpis Bottomland vs. Upland Pine 0.97 15 <0.001 Upland Pine vs. Ecotone 0.95 15 <0.001 Bottomland vs. Ecotone 0.91 15 <0.001 C. stellifer Bottomland vs. Upland Pine 0.77 15 <0.001 Upland Pine vs. Ecotone 0.38 15 0.008 Bottomland vs. Ecotone 0.44 5 0.003

Figure 2-3. Comparison of abundance of 10 Culicoides spp. in A) total trap counts of all physiological status, B) bloodmeals on white-tailed deer, and C) aspirations from white-tailed deer. These data were recorded from a big game preserve in Gadsden County, FL from July 2015 through September 2016.

CHAPTER 3 VERTICAL STRATIFICATION OF CULICOIDES BITING MIDGES AT A FLORIDA BIG GAME PRESERVE2

The distribution of vectors in the environment has a strong impact on interactions with host animals, the efficacy of vector control activities, and the transmission of vector-borne pathogens. Investigations into the vertical distribution of arthropods have found that many taxa stratify by species, life stage, and physiological status (Ulyshen

2011), a phenomenon that has been demonstrated in a variety of medically important dipteran families, including the Culicidae, Simuliidae, Ceratopogonidae, Muscidae, and

Psychodidae (subfamily Phlebotominae) (Cortez et al. 2007, Ĉerný et al. 2011, Maguire et al. 2014). In addition, field studies from Connecticut, USA, found that West Nile virus minimum infection rates were significantly greater in two vector mosquito species in canopy-level traps, compared to traps at ground level (Anderson et al. 2004). Monitoring activities for the detection and prevention of arboviral outbreaks are most effective when optimized trapping protocols are employed. Despite this, many studies conducted on arboviral vectors focus trapping efforts within 1.5m of the ground without investigating optimal trap heights for specific vector species, potentially missing insects of medical and veterinary importance that are present at higher vertical strata.

Culicoides Latreille biting midges (Diptera: Ceratopogonidae) are important vectors of pathogens of primarily veterinary importance, including epizootic hemorrhagic disease virus (EHDV), bluetongue virus (BTV), and African horse sickness virus (AHSV)

(Mellor et al. 2000). They are also competent vectors of human pathogens in the New

World, including Oropouche virus (Romero-Alvarez and Escobar 2017) and the filarial

2 Reprinted with permission from Springer Nature 52 nematode Mansonella ozzardi (Ben-Chetrit and Schwartz 2015). The status of

Culicoides as vectors of over 50 pathogens worldwide (Mellow et al. 2000) warrants developing optimized monitoring strategies not only in a horizontal plane, but also in vertical space.

Several studies have examined the use of vertical space by Culicoides species worldwide (Henry and Adkins Jr. 1975, Veras and Castellon 1998, Venter et al. 2009b).

Height based differences in total abundance, physiological status, and species composition have been recorded in Culicoides (Venter et al. 2009b). In Africa, abundance and physiological status varied with height differences of less than 1m

(Venter et al. 2009b). The human pest, Culicoides furens Poey, was collected in great abundance in the canopy in South Carolina, USA, but collection heights depended on the habitat from which they were collected (Henry and Adkins Jr. 1975). Thus, height at which Culicoides are trapped affects estimates of abundance, diversity, and species composition.

While the precise biological rationale for Culicoides exploiting the forest canopy is not known, one hypothesis is that midges use the same vertical strata as their preferred host taxa (Swanson et al. 2012). By this reasoning, hematophagous insects that primarily parasitize large mammals would be collected in greater abundance near the ground (Service 1971) compared with those insects that prefer to feed on avian hosts, especially during host-seeking physiological stages — nulliparous and parous females (Bennett 1960, Service 1971). These associations have led to the proposal that host preferences can be inferred through the collection of hematophagous insects at different vertical strata (Swanson and Adler 2010). Other research seemingly indicates

53 that midges are simply feeding on whichever host is most abundant at their preferred vertical strata, and therefore the elevation of feeding determines host affiliation (Tanner and Turner 1974, Henry and Adkins Jr. 1975).

Understanding the ecology of nuisance and vector species is an important step in developing management and maintenance plans. This information is limited for many

Florida Culicoides species, including species that have been implicated as vectors of hemorrhagic diseases such as EHDV and BTV. The present research aimed to investigate whether diverse Culicoides species are encountered in large numbers in tree canopies, whether physiological statuses of midges differ in those captured at ground versus canopy level, and whether blood-meal analysis could be used to assess the hypothesis that hematophagous Culicoides spp. are caught in highest abundance at the level of their preferred host class.

Methods

Field Collections

Trapping was conducted at four sites in 2016 and ten sites in 2017 (Figure 3-1).

At each trap site, Centers for Disease Control and Prevention (CDC) miniature light traps (Model 2836BQ, BioQuip Inc., Rancho Dominguez, CA) with LED black light arrays (Model 2790V390, BioQuip Inc.) were set up at two heights. No additional attractants were used for the traps. One trap was hung on a steel shepherd's hook with intake at 1.37 m and represented the “ground” trap during both study years. At each site a “canopy” trap was operated, which was elevated to approximately 6.00m (2016) or

9.00m (2017). Trap heights were changed between years in order to sample the midge community at multiple heights. Canopy traps were hoisted into tree canopies using a

54 rope and pulley system and were placed in trees representing the dominant cover type in the stand (hardwood or pine). Out of the ten canopy traps on the property, six traps were located in stands composed primarily of hardwood tree species and four traps were located in majority pine stands (Figure 3-1). Once trap sites were established, locations were not changed within years. Ropes were placed in trees using a “Chuckit!” brand tennis ball thrower (Doskocil Manufacturing Company, Arlington, TX) with ropes attached to the tennis balls. Ropes were subsequently dragged over the branch until the pulley was at the proper height and the rope was then secured to the tree trunk.

Trapping was conducted one night per week for an 11-week period in 2016 (44 total trap nights per height) and a 12-week period in 2017 (120 total trap nights per height) during late spring and summer (June-August) each year. Data collected previously on this property indicated that midge abundances should be high during this trapping period.

In addition to light traps, non-attractive “funnel suction traps” (Figure 3-2) were used to determine the direction of travel (ascending or descending) of midges in the canopy. These traps also served as a control to determine whether the Culicoides collections in the forest canopy were solely due to the presence of an attractive light drawing midges upwards. Sampling with funnel suction traps was conducted during the summer of 2017. The traps were constructed of a large funnel of fine mesh that terminated at the intake of a suction trap (CDC miniature light trap with bulb removed).

The funnel was 1m in diameter at its opening and constructed of plastic rings with no- see-um netting (largest mesh opening 0.6 mm). The CDC trap (bulb removed) served to deposit insects into an attached 50mL conical tube containing 95% ethanol. At each

55 site, one funnel suction trap was placed to collect insects as they descended from the canopy (Figure 3-2, at right) and the other faced down to collect insects as they moved upwards (Figure 3-2, at left). Traps were connected to 6V battery sources that were replaced every other day, allowing traps to run continuously throughout the week.

Collection tubes were replaced weekly. Funnel suction traps were located at least 15m from all light traps to avoid interactions between traps.

Laboratory Methods

Culicoides were sorted from bycatch and identified to species using morphological keys (Blanton and Wirth 1979). Midges were determined to be male or female and physiological status of all female midges was categorized as nulliparous, parous, gravid, or blood-fed. Parous midges were recognized by the presence of accumulated red pigmentation on the abdominal cuticle (Dyce 1969, Akey and Potter

1979). Due to difficulty in differentiating nulliparous and parous Culicoides venustus

Hoffman, these two physiological statuses were combined into a single group termed

“unfed”.

Blood-meal analysis was conducted by polymerase chain reaction (PCR) to determine host use of blood-fed midges. DNA was extracted using Instagene Matrix

(Bio-Rad Laboratories, Inc., Hercules, CA) using the following extraction protocol. Each blood-fed individual was homogenized with 10µL NaCl in a 1.5mL microcentrifuge tube using a pestle. Room-temperature Instagene (200µL) was added to each tube and vortexed. Tubes were incubated at 98°C for ten min and subsequently centrifuged at

3099× g for 5 min after which the supernatant was collected into a clean 1.5mL tube.

Extracted DNA was then amplified by PCR using three primer sets and cycling conditions as described by Blosser et al. (2016). All samples were initially run using a mammalian and amphibian specific primer set targeting the 16S region (F: 5'-CTC CAT

AGG GTC TTC TCG TCT T-3'; R: 5'-GCC TGT TTA CCA AAA ACA TCA C-3'). PCR protocol for this primer set was 94°C for 2 min and 35 cycles of: 30 s at 94°C, 30 s at

57°C and 1 min at 72°C. Samples were subsequently run using reptile specific 16S region primers (F: 5'-CTG ACC GTG CAA AGG TAG CGT AAT CAC T-3'; R: 5'-CTC

CGG TCT GAA CTC AGA TCA CGT AGG-3') using the following protocol: 2 min at

94°C and 35 cycles of 30 s at 94°C, 30 s at 62.5°C and 1 min at 72°C. Finally, primers targeting the avian cytochrome b oxidase (F: 5'-GGA CAA ATA TCA TTC TGA GG-3';

R: 5'-GGG TGG AAT GGG ATT TTG TC-3') were used with the following protocol: 2 min at 94°C and 35 cycles of 30 s at 94°C, 30 s at 60°C and 1 min at 72°C. All protocols included a negative control (molecular grade water) and a positive control (positive controls used included house mouse (mammalian), green anole (reptilian), and chicken

(avian)).

Agarose gel electrophoresis was used to determine presence of PCR amplicons, using a 1% agarose gel (100V for 35 min). Amplicons were sent to a commercial laboratory for Sanger sequencing (Eurofins Genomic, Louisville, Kentucky, USA).

Sequences were then compared with available sequence information in the GenBank database (National Institutes of Health: National Center for Biotechnology Information) using BLAST (Basic Local Alignment Search Tool). Samples with identity matches of

≥95% and query coverages of 75% or higher were considered a host species match.

Data Analysis

Statistical analyses were restricted to those species with greater than 100 individuals collected per sample year. Individual species abundance was analyzed using negative binomial regression analysis to account for overdispersion of data and unequal mean and variance (Dinesh et al. 2008, Rigot et al. 2013). Negative binomial regression also was used to investigate differences in species abundance between traps located in stands composed of primarily hardwood or pine trees during the 2017 sampling year. Pearson’s chi-square test of independence was used to analyze whether relative proportions of each physiological status for each species were dependent upon trap height. Blood-meal analysis results also were analyzed using Pearson’s chi-square analysis to test for independence of host class use (Mammalia and Aves) between the two trap heights (canopy and ground). All analyses were performed using R software (R

Core Team 2016, R 3.3.2 software, Vienna, Austria).

Results

Total Collections

For the 2016 trapping season, four trap comparisons were run for 11 weeks resulting in 44 trap nights per height (or 88 trap nights total). This resulted in a total of

4,769 midges collected on the big game preserve. Out of the total, 1,996 (41.85%) were collected in ground traps and 2,773 (58.15%) were collected in canopy traps. These collections comprised 14 Culicoides species (Table 3-1). We observed that most species (9/14) were collected in greater abundance in the forest canopy. The species for which at least 100 specimens were collected in 2016 included Culicoides

58 haematopotus Malloch (n=914), Culicoides stellifer Coquillett (n=3,363), and C. venustus (n=418).

In 2017, the canopy trap height was changed to 9.00m and the trapping effort was increased to ten sites run for 12 consecutive weeks resulting in 120 trap nights per height (240 trap nights total). This effort collected 26,928 individuals from 22 Culicoides species. The distribution between trap heights was 2,206 (8.19%) in ground traps and

24,722 (91.81%) in canopy traps. A majority of species collected during this year were found in the forest canopy (17/22 species; Table 3-1). The species for which greater than 100 midges were collected included Culicoides arboricola Root (n=217), Culicoides biguttatus Coquillett (n=127), Culicoides debilipalpis Lutz (n=744), C. haematopotus

(n=1,409), Culicoides insignis Lutz (n=142), C. stellifer (n=21,983), and C. venustus

(n=2,116).

Ground versus Canopy Use by Abundant Species

Species for which at least 100 specimens were collected were analyzed to look for intraspecific height preferences. In 2016, none of the most abundant species showed a significant difference between canopy and ground traps (C. haematopotus: df=45, P=0.673; C. stellifer: df=83, P=0.249; C. venustus: df=59, P=0.052). In 2017, the species with significantly greater abundance in canopy collections included C. arboricola (df=96, P<0.001), C. biguttatus (df=42, P<0.001), C. debilipalpis (df=156,

P<0.001), C. haematopotus (df=192, P<0.001), C. insignis (df=42, P<0.001), and C. stellifer (df=226, P<0.001). There was no significant difference between ground and canopy collections for C. venustus (df=58, P=0.052) in 2017. No Culicoides spp. were captured in significantly greater numbers in ground traps.

Funnel Suction Traps

Funnel suction traps were operated throughout the summer of 2017 and collected 43 biting midges. Out of 43 individuals, 37 (86%) were collected ascending into the forest canopy while six (13.95%) were collected descending towards the ground

(Figure 3-2). Five blood-engorged Culicoides were collected, four of which were successfully identified including two descending from the forest canopy: one C. haematopotus with a northern cardinal (Cardinalis cardinalis) blood meal and one C. stellifer with a white-tailed deer blood meal. The other two blood-engorged midges, one

C. debilipalpis and one C. stellifer, were ascending into the forest canopy with white- tailed deer blood meals. The blood meal from one blood-engorged C. stellifer was not identified. The majority of midges collected moving into the forest canopy (52.5%) were nulliparous females.

Habitat Associations of Abundant Species

Total abundance per habitat type for the seven abundant Culicoides species was

12,725 in hardwood stands and 14,013 in pine stands. Binomial regression analysis indicated that the interaction of trap height and habitat type was not significant for C. arboricola (df=47, P=0.996), C. biguttatus (df=27, P=0.403), C. debilipalpis (df=47,

P=0.076), C. insignis (df=35, P=0.181), and C. stellifer (df=47, P=0.637). A significant interaction of trap height and habitat type was identified for C. haematopotus (df=47,

P<0.001) and C. venustus (df=47, P=0.042) (Figure 3-3). For both of these species, greater proportions of midges were found in the ground collections in hardwood dominated habitats compared with pine habitats.

Physiological Status

During 2016, there were significant differences between proportions of physiological statuses for all three species analyzed: C. haematopotus, C. stellifer, and

C. venustus. In 2017, significance was found for C. arboricola, C. haematopotus, C. insignis, C. stellifer, and C. venustus. Physiological status proportions for C. biguttatus and C. debilipalpis were not significant and, therefore, independent of trap height. Chi- square results are reported in Table 3-2 and physiological status data for three of the most abundant species, C. haematopotus, C. stellifer, and C. venustus, are shown in

Figure 3-4. In almost all cases, greater abundance of each physiological status was found in canopy traps. Within trap heights, however, variable distributions of each physiological status were observed (Table 3-3). The proportion of gravid females was higher in ground traps than canopy traps for all species with significant chi-square results for both study years. The proportions of all nulliparous females, with the exception of C. haematopotus in 2016 and unfed C. venustus, were greater in canopy traps than ground traps for both years.

Blood Meal Analysis

During the 2016 sampling season, 222 total blood-engorged midges were collected, with 119 (53.6%) yielding identity matches in BLAST of 95% or greater.

Canopy and ground collections accounted for 94 (79%) and 25 (21%) specimens, respectively. Two avian blood meals were recorded from canopy level traps, one by C. arboricola on a black vulture (Cathartes aura) and the other by C. haematopotus on a red-eyed vireo (Vireo olivaceus). There were no avian blood meals taken by midges at ground level. The great majority of mammalian blood meals were from midges collected

61 from the canopy traps (n=92, 79.3% of total), while only 20.7% of mammalian blood meals were taken from ground traps (Figure 3-4). Out of the 92 mammal blood meals taken from midges from canopy traps, 82 were from large ground-dwelling ungulate species and nine were from humans. In addition, a blood meal was taken by Culicoides crepuscularis Malloch from a gray squirrel (Sciurus carolinensis), that could have been present in the canopy at the time of blood-feeding. Chi-square analysis indicated that the proportion of blood meals collected from each host class (Mammalia and Aves) was independent of trap height during the 2016 season (χ2<0.001, df=1, P=0.577).

A total of 790 blood-engorged female Culicoides were collected during the 2017 season, with 693 of these (87.7%) successfully matched to a host at the 95% identity match level. One species, C. stellifer, constituted 70% (n=482) of the total blood meals in the dataset. The second greatest number of blood meals analyzed came from C. debilipalpis (20% of blood meals; n=140). All other Culicoides species for which blood meals were analyzed constituted less than 5% of the total dataset. Canopy traps accounted for the vast majority of collections (89.2%, n=618 blood meals), while ground traps only collected 75 blood-engorged females (10.8%). A total of 611 mammalian blood meals and seven avian blood meals were observed from midges collected in canopy traps while 73 mammalian and 2 avian blood meals were recovered from midges collected at ground level (Figure 3-4). Mammalian blood meals from canopy collected midges were mostly (98.7%) from large ground-dwelling mammals except for one C. stellifer blood meal from a raccoon (Procyon lotor), one blood meal by C. debilipalpis on a Peromyscus sp. mouse, and six blood meals from eastern gray squirrel

(four by C. debilipalpis, one by Culicoides paraensis Goeldi and one by C. stellifer).

Host class was independent of trap height (χ2=0.342, df=1, P=0.268).

Discussion

The results of this study indicate that adult Culicoides readily move into and exploit the forest canopy during each physiological state. Through the use of light baited traps, we identified that most Culicoides were collected in the forest canopy both years, although a much greater proportion of midges were collected in the canopy during the

2017 sampling year. The relative dissonance observed between study years may be associated with the difference in total trap nights between years (44 per height in 2016 compared with 120 per height in 2017). This pattern could also be associated with the height at which canopy traps were placed between years. In 2016, traps were located at

6 m height, much lower than traps in 2017 that were placed at 9 m. Previous studies have indicated that in hardwood stands, leaf area index (LAI) is generally highest in the upper third of the forest canopy (Vose et al. 1995). This could have led to increased cover for insects higher into the forest canopy.

Our analysis identified many species found in significantly greater abundance in the forest canopy than at ground level. While the impetus for this trend is not fully understood, this finding has implications for the collection and quantification of local

Culicoides populations of interest. Two of these species, C. debilipalpis and C. stellifer, are both suspected vectors for EHDV (Mullen et al. 1985a), and C. insignis is a confirmed BTV vector (Tanya et al. 1992). Quantifying abundance of these species is an important aspect of understanding the epidemiology of these veterinary pathogens.

Many ongoing studies investigating Culicoides of veterinary importance place traps at

63 low heights or fail to mention trap height placement (Veggiani et al. 2011, Brugger et al.

2016). The data collected during the present study indicate that this practice may lead to biased sampling and inaccurate approximations of vector abundance.

While the majority of midges in both years were collected in the forest canopy, exceptions did occur and species that were more frequently captured at the ground level included Culicoides beckae Wirth & Blanton and C. biguttatus in 2016 and

Culicoides loisae Jamnback, Culicoides scanloni Wirth & Hubert, and Culicoides torreyae Wirth & Blanton in 2017. Few specimens were caught for these species (10 total individuals), which prevented meaningful statistical analysis. As of now, it is unclear whether these species are selecting lower vertical strata or if this observation is a by-product of low sample sizes. Additional studies are warranted to better understand the ecology of rare Culicoides species. It is important to note that two of these species,

C. loisae and C. torreyae, are not strongly hematophagous and likely do not rely solely on blood meals for egg production (Blanton and Wirth 1979). For this reason, it is unlikely that low flight habits are associated with searching for host animals.

While other studies investigating canopy use by Culicoides spp. and other hematophagous Diptera have used light- or CO2-baited traps (Henry and Adkins Jr.

1975, Venter et al. 2009b, Swanson and Adler 2010), there are concerns that such collections are biased by the presence of light and attractive baits. For this reason, the present study supplemented light trap collections with unbaited funnel suction traps that were run continuously throughout each trap week, which successfully collected midges as they moved between the canopy and ground level without the use of an attractive bait. Further, our funnel traps only covered roughly 4m2 of the property’s total area

(200ha). If our findings are extrapolated over the full area of the preserve, this indicates a substantial movement of Culicoides (2.1 million individuals) through vertical space during the study period. It is important to note, however, that while funnel traps collected various physiological statuses from a variety of species, no gravid individuals were collected using this method. It is unclear whether this is due to relatively low sample sizes collected with this method or if gravid individuals are not transitioning between strata as often as light traps indicated. Furthermore, it is possible the fine mesh used for collecting Culicoides in funnel suction traps were viewed by the midges as obstacles or enclosures rather than open space for transition. The low sample sizes collected could be partially due to the midges actively avoiding entering this enclosed space.

Forest canopies dominated by hardwood versus pine trees are fundamentally different habitats due to differences in crown density, leaf area index, and retained moisture (Parker 1995, Liames et al. 2018). For two species, C. haematopotus and C. venustus, a significant interaction between trap height and habitat type was found, which indicated that for these two species, habitat may be an important factor in how midges stratify vertically between the ground and canopy. Both C. haematopotus and C. venustus are ground-breeding species, and as such, it is unlikely that these habitat preferences are due to increased prevalence of tree-associated breeding habitats such as tree-holes (Blanton and Wirth 1979). Despite lower retained moisture and crown density associated with pine stands (Parker 1995), a greater abundance of Culicoides was detected from pine habitats overall, although this result was not significant. This result could be due to the lower leaf area index of pine stands (Liames et al. 2018)

65 leading to higher visibility of the black light in these canopies compared to hardwood canopies.

Height did play a role in the proportion of the different physiological statuses collected for certain Culicoides species, suggesting that midges were using vertical space differently based on their physiological status. In each species where chi-square results indicated significantly disproportionate physiological status values, gravid individuals comprised a greater proportion of the total abundance in ground traps than in canopy traps. Despite this, overall abundance of total midges and gravid midges was greater in the canopy compared with ground traps. This is in agreement with the results of Venter et al. (2009b) that found greater abundance of gravid C. imicola as trap height was increased from 0.6 m to 2.8 m, while the relative proportion of gravid individuals decreased as height increased. Gravid individuals collected in greater relative abundance in ground traps may be using lower strata to search for suitable oviposition sites. This would be in agreement with research on mosquitoes that found more gravid

Culex pipiens in ground catch basin traps than in canopy traps, which was attributed to gravid females seeking out oviposition sites (Anderson et al. 2006). For three of the midge species, C. haematopotus, C. stellifer, and C. venustus, greater proportions of gravid individuals in ground traps can be explained by their larval development sites, which are typically moist substrate along water bodies, muddy puddles and muddy hoofprints (Blanton and Wirth 1979). This explanation does not account for container

(treehole) breeders, including C. arboricola and C. debilipalpis, indicating that other factors, such as body mass, may be contributing to the use of lower strata by gravid females. In addition to tree-holes, C. arboricola has been documented ovipositing in

66 woody debris and tree stumps (Blanton and Wirth 1979). Assuming low availability of tree-holes, this species may seek out these oviposition sites in woody debris on or near the ground. Greater proportions of unfed C. venustus and nulliparous midges often were found in the forest canopy, although the reason for this pattern is more challenging to parse out. This trend could be related to increased sugar feeding options, locating more suitable conditions for survival, or blood meal availability.

One of the objectives of the present study was to determine whether midges collected in the forest canopy were collected at the height representative of their preferred host range. Our analyses indicated that there was no significant association between the height at which blood-engorged midges were collected and the class of vertebrate they fed upon during either study year. These data do not agree with the hypothesis that collection height can be used to determine host preferences (Swanson and Adler 2010). This hypothesis is further challenged by the presence of canopy- dwelling mammals such as eastern gray squirrels (Sciurus carolinensis) in blood meals of midges collected in ground traps. For these reasons, using trap height to make inferences on host breadth is likely not an effective method of inferring host range and alternative methods should be used when possible. The distribution, number, and type of sensory pits on the antennal flagellomeres and maxillary palps of Culicoides has been proposed as another method to determine host preference (Braverman and Hulley

1979, Isberg et al. 2013). This hypothesis was not tested in the present study due to low blood-fed Culicoides species diversity but is an important topic for future studies. While our data do not support the hypothesis that collection height determines host- preference, the large ruminant abundance on this property was artificially high

67 compared with a more natural system due to stocking and exclusion of predators. It is possible that the great abundance of large ruminants and their volatiles could have attracted more midges down from the canopy than in a more natural system where large hoofstock is not as readily available.

Few data on the distances moved by blood-engorged Diptera are available, however, Edman & Bidlingmayer (1964) demonstrated that blood-engorged females of nine mosquito species flew at least one mile after engorgement. Data on horizontal flight distance by blood-engorged Culicoides spp. are not available and is an important topic for future research. However, our data indicate that blood-engorged Culicoides are feeding on the ground and moving vertically up to 9 m into the forest canopy to digest their blood-meals.

Our finding that midges occupy the forest canopy in great abundance also indicates that current calculations of vectorial capacity may be low for some vector

Culicoides species, if trap height is not being taken into account. Garrett-Jones (1964) identified this potential pitfall as an area of concern in calculating vectorial capacity for malarial vectors: blood-engorged arthropods move and are not always collected in the areas where they originally fed. Not only was a great abundance of midges collected in the forest canopy at heights that may not typically be investigated, but a majority of blood-engorged midges in the canopy had fed upon large ungulates, including those susceptible to pathogens such as EHDV and BTV. One component of the vectorial capacity equation is the biting rate of the vector on the affected host. If blood-meal analysis is being used to investigate biting rates for vectorial capacity calculations, biting rate could be vastly underestimated by only sampling midges at ground level and

68 underestimating a large proportion of midges that feed on ungulates on the ground and then ascend into the canopy level.

While our data demonstrate that large numbers of Culicoides can be encountered in tree canopies, the reason for this vertical movement is unknown, but it appears that this movement is likely not associated with host use and blood-feeding behavior. Other potential explanations for this movement include locating an optimal microclimate for survival or finding floral and extra-floral nectaries in trees. The microclimate of the forest canopy is variable at different heights, with the higher canopy receiving more radiation, light and heat, and experiencing greater circadian fluctuations than the lower canopy, which maintains more humid, cooler conditions and more circadian stability (Parker 1995). Future studies should investigate whether microclimate explains vertical habitat selection by Culicoides species.

The presence of attractive sugar sources including floral and extra-floral nectaries could entice midges into the tree canopy. A study conducted by Kaufmann et al. (2015) in Europe found that around 80% of Culicoides collected at a farm were fructose-positive indicating frequent sugar feeding across all physiological stages. This study found an increase in longevity of midges with access to ample sugar in addition to blood meals; however, Kaufmann et al. (2015) also found that increased longevity was not associated with the presence of extrafloral nectaries specifically. Mullens (1985) found that sugar-feeding was uncommon in C. sonorensis (as C. variipennis), with only up to 24% sugar-feeding detected in parous individuals. Culicoides represent a diverse assemblage of species with different characteristics and complex behaviors, so we

69 speculate that vertical stratification may be linked to multiple driving forces during each physiological status.

While this research has added valuable information on important Culicoides species present in Florida, many questions remain unanswered. It is unclear how frequently midges travel between the ground and canopy and at what time of day these transits are most likely to occur. A 1955 study by Snow identified diel movements of C. haematopotus and Culicoides guttipennis Coquillett, which were mainly in the forest canopy throughout the day and descended to ground level to feed during the evening

(Snow 1955). This would be an excellent future direction for research into this phenomenon. Future studies also should trap along forest edges in addition to within the forest to look for abundance differences between these two sites, which have different canopy characteristics (Sherich et al. 2007). This could have implications for the spread of Culicoides-borne diseases to farms that border forest edges. The biggest question that remains is the biological imperative driving midges into the forest canopy and whether this imperative is maintained through different physiological statuses.

The present study has demonstrated a great abundance of Culicoides biting midges present in forest canopies on a big game farm in Florida, USA. The abundance of C. arboricola, C. biguttatus, C. debilipalpis, C. haematopotus, C. insignis, and C. stellifer was significantly higher in the canopy than in ground collections in one study year. Physiological status was significantly dependent on trap height for C. haematopotus, C. stellifer, and C. venustus in 2016, and for C. arboricola, C. haematopotus, C. insignis, C. stellifer, and C. venustus in 2017. Despite speculation that host breadth could be inferred through height at collection, 98.6% of total blood-

70 engorged midges collected in the canopy were found to have fed upon ground-dwelling mammals. These data not only add to our knowledge on these understudied Culicoides species, but also provide valuable insight into opportunities for surveillance and control of biting midge species of veterinary importance in Florida.

Figure 3-1. Map depicting property boundaries, habitat classes, and trap sites at a big game farm. Big game preserve located in Gadsden county, Florida. Each trap site (purple hexagons; sites with hexagons with central black dots were run during both 2016 and 2017, sites without black dots were only run during 2017) had two Centers for Disease Control and Prevention (CDC) miniature light traps present, one at ground level and one in a tree canopy. Funnel traps for passive collection of Culicoides moving up into and down from the canopy also were located at trap sites 2 and 3 in 2017.

Figure 3-2. Funnel suction traps, biting midges captured and their sex and physiological status. The trap at left collected insects as they ascended into the canopy and the trap at right collected insects as they descended from the canopy. Funnel traps were hung halfway between the canopy and ground traps (6m) at two sites on the Gadsden county property for the duration of the study period in 2017

Table 3-1. Total Culicoides collected per species and height during the 2016 and 2017 sampling period. Total trap nights were 88 (44 per height) for 2016 and 240 (120 per height) for 2017. Species for which at least 10 individuals were collected are showna 2016 2017 Species Canopy Ground Canopy Ground Total C. arboricola 11 1 202 15 229 C. beckae 0 1 43 0 44 C. bickleyi 0 0 23 4 27 C. biguttatus 1 2 118 9 130 C. crepuscularis 2 0 38 0 40 C. debilipalpis 31 12 689 55 787 C. haematopotus 404 510 1,218 191 2,323 C. insignis 0 0 122 20 142 C. nanus 1 1 27 0 29 C. pallidicornis 0 0 14 2 16 C. spinosus 1 1 3 7 12 C. stellifer 2,013 1,350 20,231 1,752 25,346 C. venustus 300 118 1,976 140 2,534 Other species 9 0 18 11 38 Total 2,773 1,996 24,722 2,206 31,697 a Other species excluded from the table include C. baueri (n=4), C. furens (n=2), C. guttipennis (n=9), C. hinmani (n=6), C. loisae (n=2), C. ousairani (n=4), C. paraensis (n=2), C. scanloni (n=1), C. torreyae (n=4), and C. villosipennis (n=4)

Figure 3-3. Vertical stratification of Culicoides species, by habitat. Species shown include a) Culicoides arboricola, b) Culicoides biguttatus, c) Culicoides debilipalpis, d) Culicoides haematopotus, e) Culicoides insignis, f) Culicoides stellifer, g) Culicoides venustus, and h) total Culicoides females. Bars represent average females collected by Centers for Disease Control and Prevention miniature light traps for which greater than 100 total individuals were collected. Asterisks denote a significant interaction between habitat and trap height at α=0.05 (negative binomial regression) in 2017.

Table 3-2. Pearson’s chi-square results for distributions of sex and physiological status of Culicoides in ground and canopy traps on a big game preserve, Gadsden Co. FL, USA. Female physiological status was categorized as nulliparous, parous, gravid, and blood-fed. Distribution of males also was included. Nulliparous and parous counts were combined for Culicoides venustus due to an inability to reliably determine parity. Males were excluded from the chi- square analysis for Culicoides biguttatus since no male individuals were collected at either height. Year Species Χ2 df P-value 2016 C. haematopotus 26.991 4 <0.001 *** C. stellifer 266.050 4 <0.001 *** C. venustus 10.656 3 0.014 ** 2017 C. arboricola 81.905 4 <0.001 *** C. biguttatus 3.396 3 0.335 C. debilipalpis 4.660 4 0.324 C. haematopotus 81.905 4 <0.001 *** C. insignis 10.050 4 0.04 * C. stellifer 107.830 4 <0.001 *** C. venustus 13.273 3 0.004 ** *P<0.05, **P<0.01, ***P<0.001

Figure 3-4. Vertical distribution, sex and physiological status, and host use of Culicoides on a private big game preserve in Gadsden Co. FL in 2016 and 2017. Parous and nulliparous groups were combined to account for difficulties in identifying parity in Culicoides venustus. Treemaps represent the results of blood-meal analysis for that species in the ground and canopy traps each year. Only the three most commonly collected species are represented. The proportions of sex and physiological statuses were significantly different in canopy versus ground traps for all species shown at alpha<0.05.

Table 3-3. Sex and physiological status for ground and canopy collected Culicoides in 2016 and 2017 with proportion in parentheses. Physiological statuses of females are shown are nulliparous (NP), parous (P), gravid (GR), blood fed (BF), and unfed (UF). Males (M) also are shown. Only species with >100 individuals collected are presented. Nulliparous and parous individuals of Culicoides venustus were combined into a single group, unfed (UF), due to difficulty seeing pigmentation on the abdomen of parous females.

Year Species Height NP P GR BF M UF† Total 45 285 38 3 33 2016 C. haematopotus Canopy - 404 (0.11) (0.71) (0.09) (0.01) (0.08) 85 277 85 2 61 Ground - 510 (0.17) (0.54) (0.17) (0.01) (0.12) 483 717 377 158 278 C. stellifer Canopy - 2,013 (0.24) (0.36) (0.19) (0.08) (0.14) 277 411 563 41 58 Ground - 1,350 (0.21) (0.30) (0.42) (0.03) (0.04) 121 2 19 158 C. venustus Canopy - - 300 (0.40) (0.01) (0.06) (0.53) 68 1 4 45 Ground - - 118 (0.58) (0.01) (0.03) (0.38) 128 38 8 3 25 2017 C. arboricola Canopy - 202 (0.63) (0.19) (0.04) (0.02) (0.12) 2 1 1 8 3 Ground - 15 (0.13) (0.07) (0.07) (0.53) (0.20) 18 61 37 2 C. biguttatus Canopy 0 - 118 (0.15) (0.52) (0.31) (0.02) 1 6 2 Ground 0 0 - 9 (0.11) (0.67) (0.22) 180 132 77 159 141 C. debilipalpis Canopy - 689 (0.26) (0.19) (0.11) (0.23) (0.21) 11 14 7 8 15 Ground - 55 (0.20) (0.26) (0.13) (0.15) (0.27) 540 323 183 25 147 C. haematopotus Canopy - 1,218 (0.44) (0.27) (0.15) (0.02) (0.12) 44 55 64 3 25 Ground - 191 (0.23) (0.29) (0.34) (0.02) (0.13) 41 18 11 11 41 C. insignis Canopy - 122 (0.34) (0.15) (0.09) (0.09) (0.34) 6 6 5 3 Ground 0 - 20 (0.30) (0.30) (0.25) (0.15) 7,790 7,218 2,953 535 1,735 C. stellifer Canopy - 20,231 (0.39) (0.36) (0.15) (0.03) (0.09) 527 755 327 62 81 Ground - 1,752 (0.30) (0.43) (0.19) (0.04) (0.05) 516 29 307 1,124 C. venustus Canopy - - 1,976 (0.26) (0.02) (0.16) (0.57) 56 1 20 63 Ground - - 140 (0.40) (0.01) (0.14) (0.45) † Culicoides venustus only - not applicable

CHAPTER 4 IMPLICATING CULICOIDES STELLIFER AND CULICOIDES VENUSTUS (DIPTERA: CERATOPOGONIDAE) AS VECTORS OF EPIZOOTIC HEMORRHAGIC DISEASE VIRUS

Epizootic hemorrhagic disease virus (EHDV) is an Orbivirus of veterinary importance, which occurs worldwide where competent Culicoides Latreille (Diptera:

Ceratopogonidae) vectors exist, including Africa, Asia, Australia, Europe, North

America, and South America (Weir et al. 1997, Savini et al. 2011, Ruder et al. 2015a,

Verdezoto et al. 2018). The primary mammalian hosts affected by the virus are wild ungulates, while domestic ruminants such as cattle do not typically succumb to disease

(Savini et al. 2011). Certain serotypes and strains of this pathogen, such as the Ibaraki strain of EHDV-2, have greater pathogenicity to cattle, although outbreaks associated with this strain have been isolated (Omori et al. 1970). This is in direct contrast to the closely related bluetongue virus (BTV), which causes considerable morbidity and mortality in domestic sheep, abortions in cows, and decreased milk production in dairy cattle (Elbers et al. 2008, Worwa et al. 2010, Santman-Berends et al. 2011, Nusinovici et al. 2012, Zanella et al. 2012). Due to the economic impact of BTV, extensive research has been conducted on the pathogen and arthropod vector species for this disease system. However, EHDV has not received the same amount of research due to its relatively lower impact on economically valuable industries. Important questions remain regarding the ecology and epidemiology of EHDV in North America, including which Culicoides species are transmitting this pathogen in regions of the United States where documented vectors are absent.

Deer farming is a growing industry that is being impacted by EHDV in the United

States. While deer have historically been used for meat and musk production in New

Zealand and China, respectively, the industry has been slow to develop in other countries (McIntyre 1976, Pearse 1990, Qi et al. 2011). In the United States, deer farming is a young but growing industry. Estimates place the economic impact of deer farming in the United States at $7.9 billion annually, supporting greater than 56,000 jobs

(Anderson et al. 2017). EHDV often results in mortality of wild and farmed deer worldwide, with farmers occasionally losing up to 80% of their herd (Lee and English

2011). Although autogenous vaccines are available for use within adjacent herds, evidence indicates that commercially available EHDV vaccines do not produce a sufficient humoral response for protection (Wisely and Sayler 2016), resulting in economic losses due to the purchase of ineffective vaccines and loss of valuable animals. For this reason, the need to better understand EHDV epidemiology in the

United States is gaining priority (Purse et al. 2015).

EHDV is transmitted by small (1-3mm) hematophagous flies in the genus

Culicoides, with only one confirmed vector species in the United States, Culicoides sonorensis Wirth and Jones (Foster et al. 1977, Jones et al. 1977). This species is common and abundant throughout the western and midwestern United States west of the Missouri river (Schmidtmann et al. 2011). However, in the southeastern United

States, where EHDV persists annually, C. sonorensis is rare, as evidenced by the lack of this species in multiple large-scale Culicoides surveys (Smith and Stallknecht 1996,

Smith et al. 1996b). The low abundance of C. sonorensis indicates that alternative vector species are likely present in this region of the United States where EHDV cases have been documented. A few species have been implicated as potential vector species of EHDV in the southeastern United States based on their abundance and presence

80 near affected host species. These species include Culicoides debilipalpis Lutz,

Culicoides obsoletus Meigen, Culicoides paraensis Goeldi, Culicoides spinosus Root and Hoffman, and Culicoides stellifer (Coquillett) (Mullen et al. 1985a, Smith and

Stallknecht 1996, Smith et al. 1996b). While the abundance of these species is an important consideration when incriminating potential vectors, this factor alone cannot be used to implicate a species as a vector of EHDV. Other criteria that should be fulfilled in order to implicate a putative vector species include (1) recovering virus from field collected individuals without visible blood in the gut, (2) demonstrating that the arthropod can become biologically infected after an infected blood meal, (3) demonstrating the arthropod’s ability to transmit the virus, and (4) showing a significant association between the implicated arthropod and the affected host population (World

Health Organization 1967). EHDV transmission studies in Culicoides are lacking, due primarily to the difficulties in colonizing Culicoides species and inducing blood feeding in the laboratory (Mullen et al. 1985, Venter et al. 2005). Despite this lack of information, the impact of hemorrhagic disease (HD) on deer farmers necessitates the identification of potential vectors so that management plans can be developed and implemented.

Due to the challenges associated with conducting laboratory vector competence studies on Culicoides, field-based evidence is vitally important to incriminating vectors in a diverse community; 49 Culicoides spp. are present in Florida (Blanton and Wirth 1979,

Borkent and Grogan 2009). This study was aimed at investigating the potential vector(s) of EHDV among white-tailed deer (Odocoileus virginianus) in the southeastern United

States using available vector incrimination criteria. Detection of EHDV RNA by qRT-

PCR from field collected biting midges and deer showing signs of disease during an

81 epizootic in northern Florida was conducted to quantify field infection rates. Examining data from blood meal analysis, live animal aspiration, and midge seasonal abundance collected prior to an epizootic permitted inference on host association of implicated

Culicoides species. Over the long term, once vector species have been identified, we can begin to fill gaps in knowledge on their ecology enabling the development of more targeted approaches to biting midge control.

Methods

Culicoides Sampling and Virus Detection During EHDV Epizootic in Northern

Florida

Culicoides midges and animal tissues were collected from five deer farms with suspected EHDV cases based on clinical presentation in affected animals in northern

Florida from August-October in 2017 (Table 4-1). Coordinates of the specific farm sites are not provided per the wishes of the private landowners whose properties were sampled. Coordinates to the nearest county center were 30.7151˚N, 85.1894˚W

(Jackson), 30.1508˚N, 84.8568˚W (Liberty), 30.5563˚N, 84.6479˚W (Gadsden, two farms), and 30.4312˚N, 83.8897˚W (Jefferson). Insects were sampled using Centers for

Disease Control and Prevention (CDC) miniature light traps (Model 2836BQ, BioQuip

Inc., Rancho Dominguez, CA) baited with either incandescent yellow light bulbs or LED black light arrays (model 2790V390, BioQuip Inc., Rancho Dominguez, CA) and carbon dioxide (solid dry ice). Four traps were used per night at the Jackson, Liberty, Gadsden-

2, and Jefferson county farms and five traps were used at the Gadsden-1 site. All collections were made in response to reports of EHDV-related deer mortality beginning in September, except those at the Gadsden-1 site where trapping was ongoing and

82 predated the first EHDV-related deer death. Culicoides were collected into ethanol and transported on dry ice back to the Florida Medical Entomology Laboratory for processing. Culicoides were identified based upon external morphology of the female

(Blanton and Wirth 1979), then pooled by species, location, and date (maximum pool size of 50 individuals). Midges were also pooled by parity, with pools containing only parous and gravid females (Dyce 1969, Akey and Potter 1979). Due to difficulties in identifying parity in Culicoides venustus Hoffman, nulliparous females were included in pools. No visible blood was present in any of the midges tested for virus. Samples were stored in ethanol and identified on dry ice prior to being pooled and transferred to 2mL microcentrifuge tubes containing HyClone medium 199 with Earle’s balanced salt solution (GE Healthcare Life Sciences, Chicago, IL) for homogenization. Samples were homogenized using a TissueLyser (Qiagen, Valencia, CA) set at 19.5Hz for 3 min or using a Bullet Blender Storm 24 (Next Advance, Troy, NY) at speed four for 3 min. Viral

RNA was extracted from lysate using the QiaAmp Viral RNA Mini kit following kit protocols (Qiagen, Valencia, CA).

Viral RNA was amplified using multiplex qRT-PCR protocol for BTV and EHDV using the SuperScript III Platinum One-step qRT-PCR kit (ThermoFisher Scientific,

Waltham, MA). Reagents per sample included 2.2µL molecular grade water, 12.6µL 2X reaction mix, 1µL each 10µM BTV forward and reverse primer, 0.8µL each 20µM EHDV forward and reverse primer, 0.4µL each 10µM FAM-labelled BTV probe and 20µM

Texas Red labelled EHDV probe, 0.8µL Platinum Taq/SuperScript III reverse transcriptase mix, and 5µL RNA template (50-100ng/µL concentration). BTV and EHDV primer and probe sequences were from Wernike et al. (2015). Cycling conditions were

83 modified from Wernike et al. (2015) as follows: reverse transcription at 48˚C for 10 min, initial denaturation for 10 min at 95˚C, followed by 40 cycles of 95˚C for 15 s, 57˚C for

45 s, and 68˚C for 45 s. All reaction plates contained EHDV positive control RNA from the CDC and a molecular biology grade water negative control. The serotype of EHDV vRNA positive samples was identified in a subsequent reaction using primers and probes for EHDV-1, EHDV-2, and EHDV-6 described by Maan et al. (2017). The 25µL assay was modified from the previously published method as follows: 2x Vet-MAX-Plus

One-Step qRT-PCR (ThermoFisher Scientific, Waltham, MA) was added to the reaction mixture and 5uL of RNA was utilized as template. Amplification was carried out using an

ABI 7500 FAST system (Applied Biosystems, Waltham, MA) using slightly modified conditions as follows: 48°C for 10 min reverse transcription, followed by 95°C for 10 min, and 45 cycles of 95°C for 15 s and 60°C for 60 s.

White-tailed Deer Sampling and Virus Detection during the 2017 Epizootic

Field necropsies were performed by Cervidae Health Research Initiative technicians and veterinarians on farmed white-tailed deer that succumbed to disease following clinical signs of HD infection (edema, inappetence and lethargy) at the same five farms where Culicoides sampling was conducted (Table 4-1). Whole blood was collected and transferred in RNase-free sterile tubes to the University of Florida

Cervidae Health Research Initiative (CHeRI) for further analysis.

RNA was extracted from whole blood using the QIAamp Viral RNA mini kit

(Qiagen, Valencia, CA) according to the manufacturer’s instructions. Cycling conditions, primers, and probes for multiplex qRT-PCR detection of EHDV and BTV were the same as for the Culicoides viral RNA detection described above. The master mix reagents

84 were adjusted for use with the VetMax-Plus Multiplex One-Step RT-PCR kit (Applied,

Biosystems, ThermoFisher Scientific, Waltham, MA) and VetMAX Xeno RNA internal control (Applied Biosystems, ThermoFisher Scientific, Waltham, MA) as follows: 12.5µL

2X multiplex RT-PCR buffer, 1µL each of 10µM BTV forward and reverse primer, 0.2µL

BTV FAM labelled probe, 0.75µL each of 20µM EHDV forward and reverse primer,

0.125µL EHDV Texas Red labelled probe, 1µL Xeno RNA internal control, 0.375µL molecular grade water, and 2.5µL 10x multiplex RT-PCR enzyme. All reaction plates included positive control RNA for EHDV-1, EHDV-2, and EHDV-6 provided by the

Southeastern Cooperative Wildlife Disease Study, a molecular biology grade water negative control, and a non-template Xeno internal control. PCR protocols for serotyping white-tailed deer blood samples were identical to the protocols described above for Culicoides pools.

Statistical Analysis of Virus Data

Maximum likelihood estimates (MLE) were calculated to determine the infection rate and 95% confidence intervals at each site where positives were recovered (Chiang and Reeves 1962). This metric estimates infection rates based on probabilistic models following a binomial distribution and can be adapted for use with variable pool sizes

(Biggerstaff 2003; 2009, Gu et al. 2003). MLE was calculated using the PooledInfRate

Excel add-on developed by the CDC (Biggerstaff 2009). A Fisher’s exact test was run to analyze whether serotype was associated with midge species tested. Fisher’s exact tests also were used to compare the distribution of serotypes in deer with serotypes in midge pools at each farm. In order to determine what variables might predict recovery of virus positive pools, negative binomial regression models were fitted and the best model

85 was selected by Akaike Information Criterion (AIC) utilizing a backwards stepwise selection method. Variables included the day of the year, site, light type, presence of

CO2, abundance of C. stellifer, abundance of C. venustus, and total abundance of other species collected. The response variable was the total number of positive EHDV pools.

Light type and CO2 use were not recorded for Gadsden-2 and Jefferson sites from

September 18-20, resulting in the removal of these dates from this analysis. Fisher’s exact test and negative binomial regression analyses were run using R studio (R Core

Team 2016, R 3.3.2 software, Vienna, Austria).

Live Animal Aspiration

In order to identify host-use patterns, midges were aspirated at the Gadsden-1 site from tame white-tailed deer in pens from June 2015 through September 2016.

Collections were made at three time points weekly (dawn, midday, dusk) and were conducted on any approachable animals for 10 min per session. The aspirator was swept over the entire body during this period. Aspirator design was an acrylic tube with a computer fan powered by a 12V battery. Collections were made directly into a plastic collection cup with a wire mesh bottom and stored at -20˚C until analyzed. All midges collected were identified to species using morphological keys (Blanton and Wirth 1979).

Blood Meal Analysis

Midges were collected twice weekly at the Gadsden-1 site between July 2015 and August 2017 to quantify seasonal abundance patterns and host use. This site is largely composed of a 200-hectare big game preserve harboring a variety of Bovidae and Cervidae species, with two areas of penned white-tailed deer present. Midges were collected at 20 trap sites throughout the preserve using CDC miniature light traps with

LED black light arrays set 1.62m above the ground. Between November-March, trapping was restricted to 10 sites twice per week. At 10 trap sites, additional elevated trapping at 6m in 2016 and 9m in 2017 was conducted using traps suspended from trees.

All blood engorged midges collected were analyzed to determine blood meal origin down to the level of vertebrate species using published protocols (McGregor et al.

2018). Due to difficulty in identifying between individuals in the genus Cervus, results for the elk (Cervus canadensis), sika deer (Cervus nippon), and hybrids (Cervus canadensis x Cervus nippon) are combined into a single group called Cervus spp. In addition to looking at blood meals on the primary EHDV host, white-tailed deer, the blood meal results for elk (Cervus spp.) and fallow deer (Dama dama) were also compiled. Studies have indicated that both elk and fallow deer can be viremic carriers of

EHDV (Hoff and Trainer 1973, Gibbs and Lawman 1977).

Seasonal abundance of C. stellifer and C. venustus

During 2016, animal mortality data and midge abundance data were recorded from Gadsden-1 from January-December. Collections were made twice weekly using the same 20 CDC miniature light traps at the Gadsden-1 site that were used for blood- engorged Culicoides collection (reduced to 10 traps during the months of November-

March). Total collections of female C. venustus and C. stellifer from 2016 were compiled temporally to determine periods of high midge activity. Data on EHDV-related deer mortality at this farm also were compiled from 2016.

Results

Virus Detection in Field Collected Midges during an EHDV Epizootic

From September-October, 16 traps were set at Jackson, 20 at Jefferson, 24 at

Gadsden-2, and 32 at Liberty. Trapping took place from August-October at Gadsden-1 with 97 total trap nights during this time. In total, 19,000 biting midges (661 pools) from

19 total species were screened for the presence of EHDV RNA from five deer farms during an HD (hemorrhagic disease) outbreak in northern Florida, resulting in 20 EHDV- positive pools (Table 4-2). All positive pools were from two species, C. venustus (14 positive pools, 70%) and C. stellifer (6 positive pools, 30%) identified from four of the farms, including the farms in Liberty and Jefferson Counties and both farms in Gadsden

County (Figure 4-1). The farm in Liberty County had six total positive pools out of 265 pools tested, three of the positives were from pools of C. stellifer (MLE (maximum likelihood estimate) infection rate=0.05%, 95% CI: 0.01-0.17%) and three were from pools of C. venustus (MLE=1.50%, 95% CI: 0.41-4.08%). Gadsden-1 had two positive pools, one of C. stellifer (MLE=0.05%, 95% CI: 0.00-0.23%) and one of C. venustus

(MLE=0.73%, 95% CI: 0.04-3.70%), out of 113 tested. Gadsden-2 had two positive pools, both C. venustus (MLE=3.75%, 95% CI: 0.67-12.61%), out of 79 pools tested.

Jefferson had the most positive pools with ten positive pools, two from C. stellifer

(MLE=0.10%, 95% CI: 0.02-0.33%) and eight from C. venustus (MLE=2.78%, 95% CI:

1.42-5.29%), out of 176 tested. No other Culicoides species were positive for EHDV vRNA (Figure 4-1). Most positive pools were identified as EHDV serotype-6 (n=12, 60%;

Table 4-3). Six of the pools were identified as EHDV serotype-2 (30%). Two pools, both of C. venustus collected in Jefferson County, were typed to both EHDV serotype-6 and

88 serotype-2. None of the samples were identified as EHDV-1, the other known EHDV serotype present in Florida. Serotype was independent of midge species tested

(Fisher’s exact test P=0.351). Culicoides venustus and C. stellifer were infected with each serotype in similar proportions.

In a negative binomial regression model investigating the effect of site, day of year, light type, presence of CO2, C. stellifer abundance, C. venustus abundance, and total other species abundance on the likelihood of getting a positive midge pool, the model containing the explanatory variables of C. stellifer abundance, C. venustus abundance, and presence of CO2 was the most parsimonious with an AIC of 56.731.

The next best models for predicting the likelihood of recovering a virus positive pool were C. venustus abundance and CO2 presence (AIC=56.83) and C. venustus abundance alone (AIC=57.224). In all models, C. venustus abundance was the common significant factor at P<0.05.

In addition to the EHDV positive pools, one midge pool was positive for the presence of BTV RNA in the multiplex assay. The positive came from a pool of 42 C. stellifer collected at Gadsden-2 on September 20, 2017. BTV serotypes were not determined for this isolate. BTV was detected in three deer from Gadsden-1, seven deer from Gadsden-2, and one deer from Liberty during the Culicoides sampling period.

Animal Mortality due to EHDV

From early-September to mid-October 2017, a total of 30 HD cases attributable to EHDV serotypes 2, 6 or coinfections with both serotypes were observed at all five deer farms (Table 4-3). Gross pathology observed in most cases (23/30) included generalized edema and hemorrhages involving different tissues and organs, most

89 frequently being lung, heart, spleen, and kidney. Similarly, hemorrhages were observed on the serosal surfaces of the stomach and there were multiple appreciable hemorrhages and intravascular coagulation on the pulmonary arteries. HD was determined to be the cause of death in all cases based on a combination of clinical signs, including the peracute or acute nature of disease, gross pathology and molecular data (all cases were confirmed by detection of viral RNA in the whole blood of suspect animals). At Liberty, 15 cases were confirmed: all were caused by EHDV-6 except a single case of EHDV-2 and a single case of mixed infection (confirmed in multiple tissues) with types 2 and 6. At Gadsden-1, seven deaths caused by infection with EHDV were confirmed: two were caused by EHDV-2 infection and the remainder, all of which impacted fawns, were caused by EHDV-6. At Gadsden-2, one death caused by EHDV-2 was confirmed. At Jefferson, five deaths caused by EHDV-2 were confirmed. Four of five individuals were fawns born in 2017. At Jackson, where no positive midge pools were detected, there were two deer deaths caused by EHDV-6.

At Liberty (Fisher’s exact test P=0.292), Gadsden-1 (Fisher’s exact test P=1), and Gadsden-2 (Fisher’s exact test P=1) the serotypes identified in deer and midge samples (with C. stellifer and C. venustus positives combined) were not significantly different. This indicates that the midges were positive for serotypes in similar proportions to the EHDV positive white-tailed deer on these farms. Fisher’s exact test results for the Jefferson farm were significant indicating that the proportion of detected serotypes in midge pools was independent of the serotype composition detected from deer at this farm (P=0.026).

Live Animal Aspiration

A total of 685 individuals were collected during live-animal aspirations. Both C. stellifer and C. venustus were collected from white-tailed deer using live animal aspiration. A total of 25 C. venustus and 213 C. stellifer were collected from June 2015 through September 2016 (Figure 4-2a). In 2015, C. stellifer and C. venustus were both collected feeding on deer, starting when aspirations began in June and ending in

November for C. stellifer (n=82 individuals), and December for C. venustus (n=18 individuals). In 2016, C. stellifer was collected feeding on deer from April through

September (n=131); Culicoides venustus was collected rarely (n=7) compared with C. stellifer, with the greatest abundance collected during September (n=4). Other species in aspiration collections included C. biguttatus (n=15) and C. pallidicornis (n=373), both of which were only documented approaching deer from March-May. Culicoides debilipalpis were collected from June through September (n=56) with the greatest collection in August (n=40 individuals, 71.4% of total collected). A few individuals of C. insignis (n=2) and C. paraensis (n=1) also were collected in aspirations.

Blood Meal Analysis

Blood meal analysis data from the Gadsden-1 site from June 2015 through

August 2017 indicated that both C. stellifer and C. venustus fed upon EHDV hosts throughout the year, including during periods of active EHDV transmission. The total trapping effort resulted in the collection of 116,007 Culicoides, 3,154 of which were blood-engorged (2.7% of total collection). Blood-engorged individuals were collected from 17 different species including 2,555 blood-engorged C. stellifer and 72 blood- engorged C. venustus. Successful host matches were made for 2,060 C. stellifer

(80.63% of specimens) resulting in 848 white-tailed deer, 307 fallow deer, and 507

Cervus spp. blood meals (Figure 4-2b). Blood meal analysis was successfully performed on 63 C. venustus (87.5% of specimens) resulting in 35 white-tailed deer, 4 fallow deer, and 7 Cervus spp. host matches (Figure 4-2c). The remaining 398 C. stellifer and 17 C. venustus blood meals came from other hosts not known to be hosts of EHDV. These data indicate that both of these putative vector species feed on all three of these host species, with feeding occurring throughout the late summer and fall, established transmission periods for EHDV (Stallknecht and Howerth 2004).

Seasonal Abundance of C. stellifer and C. venustus

In 2016, collections were made two nights per week on the Gadsden-1 property with 20 traps from April-October and ten traps from January-March and November-

December. The first collections for both C. stellifer and C. venustus were made in March and continued through December. Average nightly abundance of C. stellifer was greater than 100 individuals from April through October with highest average abundance occurring in May (푥̅=860, σ=762.22). The highest average abundance for C. venustus also was observed in May (푥̅=68, σ=29.07) with abundance fluctuating between 21-43 average individuals per night from June through October. The greatest period of EHDV- associated white-tailed deer mortality was observed in September with seven deaths occurring during that month, during which on average 402 C. stellifer (σ=121.42) and 21

C. venustus (σ=13.67) females were collected per night of trapping (Figure 4-3).

Discussion

The identification of 20 EHDV-positive Culicoides pools at Florida farms where active transmission of EHDV resulting in clinical disease was occurring supported the

92 incrimination of two probable vector species in northern Florida: C. stellifer and C. venustus. Culicoides stellifer has been implicated by other studies due to its great abundance during outbreaks (Mullen et al. 1985, Smith and Stallknecht 1996, Smith et al. 1996b); however, C. venustus has not received much attention as a potential EHDV vector (Smith et al. 1996b). Despite this, C. venustus accounted for the majority of

EHDV-positive pools and had higher infection rates than C. stellifer at all farms in this study. Culicoides venustus, along with C. insignis, is a member of subgenus Hoffmania, a mostly tropical subgenus with several vector species (Mellor 1990).

Our finding that serotypes were largely equivalent between midge pools and animal samples, except for the Jefferson county farm, provides additional evidence for implicating these species as vectors of EHDV in the southeastern United States. For the

Jefferson county farm, the result that EHDV-2 was only identified in Culicoides pools and not recovered from deer samples could be attributable to a variety of causes. Wild deer populations surrounding this farm may have had a greater prevalence of EHDV-2 than farmed populations leading to the dissimilarities in serotype distribution between deer and midges on this property. Culicoides also are believed to travel on wind currents (Ducheyne et al. 2007, Kedmi et al. 2010a), which could have transported midges from an area experiencing greater EHDV-2 activity towards the Jefferson

County farm. Finally, this result could also be attributable to stochasticity due to low sample sizes of EHDV positive animals available for sampling at the time of death.

Models indicated that abundance of both C. stellifer and C. venustus as well as the presence of CO2 were the most important factors in predicting EHDV positives, which has implications for future EHDV detection studies. The use of CO2 led to

93 increased collections of midges (10,951 individuals were collected with CO2 while only

5,963 were collected without CO2), which is a trend seen in other Culicoides studies investigating the utility of CO2 use with light traps (Anderson and Linhares 1989, Venter et al. 2016). Future studies investigating Culicoides vectors should prioritize using CO2 in traps to increase the likelihood of collecting sufficient sample sizes for EHDV detection. The inclusion of C. stellifer and C. venustus abundance in this model further reinforces their vector capacity. Despite the “other species” category having many more individuals than the C. stellifer or C. venustus fields on several trapping dates, it is still the abundance of these two species specifically that predicts the presence of EHDV positives in samples.

While both C. stellifer and C. venustus were included in the most parsimonious model, our data indicate that abundance of C. venustus was the only factor between models that significantly predicted EHDV outcome with greater abundance of C. venustus resulting in an increased probability of detecting positive pools. This speaks to the likely greater significance of this species as a vector of EHDV in these areas.

Additionally, the higher total infection rates for this species at each site lends strength to this assertion. Culicoides venustus vector competence for both BTV and EHDV has been tested in one population from New York. One positive infection each for BTV and

EHDV-1 out of 141 and 38 midges, respectively, orally exposed to virus was identified in this species; however, transmission was not demonstrated (Jones et al. 1983). The authors determined that C. venustus is likely not an efficient vector in New York.

However, evidence from other vector-borne pathogen systems indicates that different populations of the same species can have variable vector competence (Tesh et al.

1976, Bennett et al. 2002). Southeastern C. venustus populations may be more susceptible to EHDV infection than the New York populations evaluated previously

(Jones et al. 1983). Further, the dissemination rate and transmission potential of C. venustus has not been determined with any populations of this species. While greater infection rates in C. venustus may indicate greater competence for the virus by this species, it is possible that the far greater abundance of C. stellifer on the landscape may overcome any dearth in competence (Mills et al. 2017).

Our finding that light type is not a significant factor in collecting EHDV positive pools of Culicoides has implications for the effect of Orbivirus infection on light perception in Culicoides. A recent study indicated that BTV infection of C. sonorensis leads to an aversion for UV light (McDermott et al. 2015). EHDV-positive pools were collected using both UV and incandescent light in the present study indicating that at least some EHDV-infected midges were attracted to UV light. However, it is possible that light aversion could be seen in some individuals resulting in a lower calculated infection rate that does not adequately represent the actual infection rate on these properties. It is also possible that the UV light aversion is limited to BTV infection or is not uniform across Culicoides species. Additional research into this topic should be pursued to better understand how both of these orbiviruses may affect Culicoides physiology.

Previous studies at the Gadsden-1 site provided evidence for the frequent use of

EHDV hosts for blood meals by C. stellifer and C. venustus. These data fulfill the criterion that an interaction in time and space between hosts and incriminated vectors be demonstrated. Blood meal analysis data revealed that both species not only feed on

95 white-tailed deer in great abundance but also feed on fallow deer and elk, species that are competent at maintaining a viremia to EHDV (Hoff and Trainer 1973, Gibbs and

Lawman 1977). Furthermore, a pattern of preference for fallow deer and elk by C. stellifer on big game preserves has been found (McGregor et al. 2018b). Currently, the roles and significance of these two ruminant species in the transmission and maintenance of EHDV are as yet unknown on big game preserves in Florida and may warrant additional studies.

While the present study aimed to determine the vectors in the southeastern

United States, the range of both C. stellifer and C. venustus could have implications for their vector status in other regions as well. Both species are known to occur north to

Canada, with C. stellifer occurring throughout the United States except the far northwestern states of Washington and Oregon and C. venustus primarily occurring east of Nebraska (Borkent and Grogan 2009). Due to the high infection rates in this study and the low abundance of C. sonorensis in the eastern United States, C. venustus may be acting as the dominant vector of EHDV throughout a large portion of this region.

The lower infection rates seen for C. stellifer and the extensive range of this species could indicate that this species is acting as a secondary vector for EHDV throughout the

United States. Virus detection studies should be pursued throughout the range of both species to better understand their role as vectors of EHDV in the United States.

Despite testing 17 additional species, including species that are abundant on the landscape and have been implicated previously, no additional species were found positive for EHDV viral RNA. Culicoides debilipalpis has been implicated multiple times in other studies (Smith and Stallknecht 1996, Smith et al. 1996b) and was present in our

96 collections from all five farms sampled, but no pools of C. debilipalpis tested positive for virus. Overall populations of C. debilipalpis were low during the 2017 EHDV outbreak, and only 28 pools were tested. Due to this low abundance and lack of virus positives, we do not believe that this species is a primary vector of EHDV in this region. Culicoides insignis, a confirmed vector of BTV (Tanya et al. 1992) also present at all five deer farms, was the second most abundant Culicoides species collected, and constituted the majority of pools from Liberty County. Despite this, no samples from this species were positive for EHDV viral RNA. A further 46 pools of C. haematopotus, the third most abundant species collected, also tested negative for EHDV viral RNA. Due to the lack of virus positives despite exhibiting great relative abundance, it is unlikely that either C. insignis or C. haematopotus are vectors of EHDV in this region. This lack of positives from abundant species, combined with our finding of multiple positives from both C. stellifer and C. venustus, also indicate that the virus positive pools detected were likely not false positives or contaminated samples and represented true infections in those

Culicoides species.

Despite the large-scale virus testing conducted, only one pool of C. stellifer tested positive for the presence of BTV viral RNA at the Gadsden-2 site. Interestingly, despite testing over 5,000 C. insignis in the present study, no BTV positive samples were detected from this species, despite it being a confirmed BTV vector (Tanya et al.

1992). Previous virus detection studies in Louisiana and Florida failed to detect BTV in

884 and 200 individual C. stellifer, respectively (Greiner et al. 1985, Wieser-Schimpf et al. 1993). While intrathoracic inoculations of BTV into C. stellifer have resulted in low infection rates, infection through viral blood meals has not been successfully

97 demonstrated with this species (Mullen et al. 1985). Culicoides stellifer may be involved in BTV transmission; however, based on the low infection rate in current and other studies, it is unlikely that C. stellifer is playing a dominant role in the transmission of this pathogen.

There were a few limitations to this study. The first was the lack of variation in trap heights for virus detection. Many Culicoides species in the Florida panhandle, including C. stellifer, C. debilipalpis, C. haematopotus, and C. venustus, are known to frequent forest canopies, often descending to take blood meals and then moving back into the forest canopy (McGregor et al. 2018a). Our collection of midges at one height may have led to underestimations of vector-positive midges in this study. Another limitation is our inability to draw inferences about EHDV transmission on the natural landscape. We do not have adequate data on EHDV prevalence or serotypes found in wild populations to determine whether the patterns observed at the deer farms are reflected in the natural ecosystems of north Florida.

The present study has identified two species, C. stellifer and C. venustus, as probable vectors of EHDV in the southeastern United States, fulfilling two of the four

World Health Organization criteria for vector recognition for both of these Culicoides species. Viral RNA was detected in field collected individuals of both species lacking blood in the gut and an interaction in time and space between the host, white-tailed deer, and the putative vector species was demonstrated. While the last two criteria, showing infection and transmission potential for Florida populations of these two species, have not been fulfilled yet, we believe it is of the utmost importance to establish the most likely vector species of EHDV in this region. Identification of the vector species

98 can lead to targeted control efforts for deer farmers in the state and direct future studies on Culicoides ecology and biology towards those species most likely to transmit pathogens.

Table 4-1. Summary of sampling locations, sampling frequency, and epizootic hemorrhagic disease virus (EHDV)-positive samples collected from dead deer in 2017. Samples were taken from four counties in Florida: Jackson, Liberty, Gadsden, and Jefferson. Trapping was conducted using Centers for Disease Control and Prevention (CDC) miniature light traps baited with either UV or incandescent light and, when available, CO2. Trapping at Jackson, Liberty, Gadsden-2, and Jefferson took place only after an initial EHDV-related mortality from September-October. August data is included for Gadsden-1 due to ongoing trapping at this location at the time of the EHDV outbreak. County Month Days Sampled Trap Nights Deer Sampled Jackson September 2 8 2 October 2 8 0 Liberty September 0 0 13 October 6 20 2 Gadsden-1 August 3 12 0 September 9 45 1 October 8 40 5 Gadsden-2 September 3 12 1 October 3 12 0 Jefferson September 3 12 4 October 5 20 1 Grand Total 44 189 29

100

Table 4-2. Total Culicoides sampled, number of pools per species and epizootic hemorrhagic disease virus (EHDV) positive pools during the EHDV epizootic in northern Florida counties (August-October, 2017). Values represent total Culicoides sampled and the number of positive pools over the total pools per site in parenthesis. Culicoides were sampled using Centers for Disease Control and Prevention (CDC) miniature light traps. Species Jackson Liberty Gadsden-1 Gadsden- Jefferson 2 arboricola 0 3 (0/3) 1 (0/1) 1 (0/1) 24 (0/6) baueri 0 2 (0/2) 0 0 40 (0/4) bickleyi 0 0 0 0 11 (0/2) biguttatus 1 (0/1) 240 (0/11) 0 0 0 crepuscularis 2 (0/2) 0 0 0 0 debilipalpis 1 (0/1) 38 (0/8) 13 (0/5) 23 (0/7) 33 (0/7) furens 0 0 0 0 648 (0/22) haematopotus 5 (0/1) 10 (0/6) 426 (0/20) 480 (0/19) 0 insignis 131 (0/8) 4,790 (0/112) 142 (0/15) 125 (0/14) 1,528 (0/43) mississippiensis 0 0 0 0 175 (0/11) nanus 0 0 0 2 (0/1) 0 pallidicornis 0 0 0 0 22 (0/3) pusillus 0 0 0 2 (0/1) 1 (0/1) scanloni 0 0 2 (0/1) 0 0 spinosus 2 (0/1) 260 (0/13) 0 0 54 (0/8) stellifer 116 (0/9) 4,016 (3/91) 2,065 (1/60) 658 (0/22) 2,032 (2/48) variipennis 0 1 (0/1) 0 0 0 venustus 4 (0/4) 215 (3/18) 140 (1/11) 60 (2/14) 454 (8/21) Total 263 (28) 9,575 (265) 2,789 (113) 1,351 (79) 5,022 (176)

101

Infection Rate (%) 5 4 3 2 1 0

C. debilipalpis Negative Pools C. furens Positive Pools C. haematopotus Infection Rate

Liberty C. insignis C. stellifer C. venustus C. debilipalpis C. furens C. haematopotus C. insignis

Gadsden-1 C. stellifer C. venustus C. debilipalpis C. furens C. haematopotus C. insignis

Gadsden-2 C. stellifer C. venustus C. debilipalpis C. furens C. haematopotus

C. insignis Jefferson C. stellifer C. venustus

0 25 50 75 100 125 Pools (N)

Figure 4-1. Epizootic hemorrhagic disease virus (EHDV) infection rate of six Culicoides spp. at five Florida deer farms during a hemorrhagic disease (HD) epizootic, 2017. Midges were collected using Centers for Disease Control and Prevention (CDC) miniature light traps. Infection rates were calculated using maximum likelihood estimates (MLE) adjusted for use with variable pool sizes. Thirteen less common species are not presented.

102

Table 4-3. Serotypes of EHDV recovered from Culicoides stellifer, Culicoides venustus, and Odocoileus virginianus (white-tailed deer) during the 2017 hemorrhagic disease outbreak. Values represent numbers of epizootic hemorrhagic disease virus (EHDV) serotypes 2 and 6 detections. Midges were collected using Centers for Disease Control and Prevention (CDC) miniature light traps (August-October, 2017). Tissues from white-tailed deer that were suspected to have died from EHDV were taken during necropsies at the same farms where Culicoides were collected. qRT-PCR was used to confirm the presence of EHDV and serotype in all samples. Serotype Species Jackson Liberty Gadsden-1 Gadsden-2 Jefferson 2 C. stellifer 0 1 0 0 0 C. venustus 0 1 1 1 2 O. virginianus 0 1 2 1 5 6 C. stellifer 0 2 1 0 2 C. venustus 0 2 0 1 4 O. virginianus 2 13 5 0 0 2 and 6 C. stellifer 0 0 0 0 0 C. venustus 0 0 0 0 2 O. virginianus 0 1 0 0 0

103

Figure 4-2. Host use of Culicoides stellifer and Culicoides venustus at the Gadsden-1 site. Host use was determined by direct aspiration from tame white-tailed deer from June 2015-September 2016 (Panel a) and PCR-based blood meal analysis from C. stellifer (Panel b) and C. venustus (Panel c) collected using Centers for Disease Control and Prevention (CDC) miniature light traps from June 2015-August 2017 at a big game preserve located in Gadsden County, Florida.

104

Figure 4-3. Seasonality of Culicoides stellifer, Culicoides venustus, and epizootic hemorrhagic disease virus (EHDV) related mortality at the Gadsden-1 farm in 2016. Collections were made using Centers for Disease Control and Prevention (CDC) miniature light traps with black light LED arrays set throughout the Gadsden-1 property from January-December (20 traps used April-October and 10 traps November-March), 2016. Mean abundance and standard deviation of midges collected per night during each month are shown. Data have been log-transformed for clarity.

105

CHAPTER 5 VECTOR COMPETENCE OF CULICOIDES SONORENSIS FOR EPIZOOTIC HEMORRHAGIC DISEASE VIRUS SEROTYPE 2 STRAINS FROM CANADA AND FLORIDA

Epizootic hemorrhagic disease virus (EHDV) is a Culicoides-borne Orbivirus causing morbidity and mortality in domestic and wild ungulates of the families Cervidae and Bovidae. Infection with this pathogen can result in decreased milk production, mouth lesions, and mortality in cattle (Kedmi et al. 2010b, Stevens et al. 2015), antler growth abnormalities and mortality in mule deer (Dubay et al. 2004, Fox et al. 2015), and considerable mortality, chronic lesions and hoof abnormalities in white-tailed deer

(Davidson 2006, Stevens et al. 2015). While this pathogen results in many adverse outcomes for affected host species, many aspects of EHDV transmission are still understudied.

Currently, three EHDV serotypes are present in the USA: EHDV-1, EHDV-2, and

EHDV-6 (Shope et al. 1955, Chalmers et al. 1964, Allison et al. 2010), with serotype determination primarily based on two outer capsid proteins, VP2 and VP5 (Anthony et al. 2009b). Within these serogroups exist a great array of strains evolving in time and space, and documentation from other pathogen systems show that even minor genetic variation between strains can have significant effects on infection and transmission in arthropod vectors and hosts (Holmes and Burch 2000, Tsetsarkin et al. 2007, Brault et al. 2018). For example, the emergence of Chikungunya in the Old World was attributable to a single amino acid change in the viral E1 envelope glycoprotein gene, which allowed Aedes albopictus Skuse (previously considered an inefficient vector) to become a primary vector in the Indian Ocean epidemic by reducing the extrinsic incubation period (period from acquisition of virus until transmission is possible) of

106

Chikungunya virus (CHIKV) and increasing its vector competence by 1000-fold

(Tsetsarkin et al. 2007, Tsetsarkin and Weaver 2011). A similar situation exists among enzootic Venezuelan equine encephalitis (VEE) viruses that usually do not amplify in equine hosts due to low viremia. The emergence of epidemic VEE is facilitated by amino acid replacements in the E2 envelope glycoprotein that enhance amplification in equines (Brault et al. 2002, Anishchenko et al. 2006, Weaver and Reisen 2010).

Significant variation between strains also has been found in the Culicoides-borne

Orbivirus African horse sickness virus (AHSV) with significant differences between rates of infection of Culicoides imicola Kieffer for multiple strains of AHSV serotypes 1, 3, and

7 (Venter et al. 2010). These relationships indicate that increased evaluation and surveillance of strain-level variations among arboviruses are warranted.

While studies documenting the effect of viral strain on infection in the arboviral vectors are lacking for EHDV, there is evidence that different strains of EHDV can have novel characteristics and produce variable outcomes for infected vertebrate hosts. One example of this phenomenon is Ibaraki disease, caused by a series of EHDV-2 strains isolated from cattle in Japan (Campbell et al. 1978). Ibaraki disease causes greater than normal morbidity and mortality in cattle, a host that is otherwise not greatly affected by other EHDV strains (Omori et al. 1970), and a 1997 outbreak of Ibaraki resulted in numerous stillbirths and abortions in cattle (Ohashi et al. 1999). Despite an extreme divergence of signs between typical EHDV-2 and Ibaraki in cattle, genetic differences between the strains were minor, resulting from a slight reassortment of viral RNA segment 2 (Ohashi et al. 2002).

107

Currently, there is only one confirmed vector of EHDV in the United States:

Culicoides sonorensis Wirth and Jones (Foster et al. 1977, Jones et al. 1977), although other species likely transmit EHDV where C. sonorensis is rare or absent. Throughout its range (western USA, northern Mexico, and southern Canada), C. sonorensis populations are likely exposed to many different strains of each EHDV serotype annually. Despite this, most vector competency studies typically assay only a single

EHDV strain and apply those results broadly, which has the potential to underestimate or overestimate risk.

The aim of this study was two-fold. The first goal was to compare the infection, dissemination, and transmission potential of C. sonorensis for two strains of EHDV-2: a reference strain of the virus collected in Alberta, Canada in 1962 (Can-Alberta strain) and a strain recently isolated from the spleen of a white-tailed deer that succumbed to the disease in Gadsden County, Florida in 2016 (FL strain). The second goal was to compare two transmission assays, honey cards and capillary assays, to investigate differences in assay efficacy with Culicoides. Honey cards were originally developed in

Australia as a method to monitor pathogen prevalence in wild mosquito populations in the absence of sentinel animals (Hall-Mendelin et al. 2010) and have since been adapted to quantify transmission (collect saliva) in arthropod vector-competence studies

(Alto et al. 2017).

Methods

Culicoides sonorensis colony

Laboratory-reared C. sonorensis from the Ausman colony maintained by the

Arthropod-borne Animal Diseases Research Unit (ABADRU) of the United States

108

Department of Agriculture (USDA, Agricultural Research Service) were used for this study. This colony was originally established in 2001 from collections of midges from

Erie, CO. Larval C. sonorensis were fed a mix of nutrient broth, bacteria inoculum from an unspecified collection site for Culicoides, and a mixture (referred to as kalf) of yeast extract, brain heart infusion medium, albumin, alfalfa powder, and a high protein supplement. Larvae were reared according to a 13:11 (L:D) photoperiod and experienced rearing temperatures ranging from 4.4˚C to 15.6˚C. Colony pupae were collected, allowed to mature, and adults were fed on 10% sucrose solution.

For this study, 1-day old Ausman C. sonorensis adults were shipped overnight to the Florida Medical Entomology Laboratory (FMEL). Midges were shipped with access to 10% sucrose solution and upon arrival were immediately aspirated into clean paper cups with no-see-um netting (greatest mesh opening 0.6mm). These cups were placed in an incubator held at 25˚C with 12:12 light:dark cycle until bloodfeeding the same day.

No sugar was provided between arrival at FMEL and blood feeding.

Viral Strains

Two viral strains were used for this study: a reference strain of EHDV-2 isolated in Alberta, Canada in 1962 obtained from the Centers for Disease Control and

Prevention (CDC) (NCBI Accession #AM74997-AM745006) and a strain isolated from the spleen of a white-tailed deer in Gadsden County, FL in 2016 (NCBI Accession

#MF688816-MF688825). Both viruses were passaged twice on African green monkey kidney (Vero) cells maintained on medium 199 with Earle’s Balanced Salt Solution

(HyClone Medium 199, GE Healthcare Life Sciences, Logan, UT) containing 10% fetal bovine serum (Atlanta Biologicals, Flowery Branch, GA), 2% penicillin streptomycin

109

(Thermo Fisher Scientific, Waltham, MA), and 0.2% Amphotericin B (Thermo Fisher

Scientific, Waltham, MA).

Bloodfeeding

Cohorts of ≤50 midges were fed EHDV-2 infected blood using a Hemotek membrane feeding system (Hemotek Ltd, Blackburn, UK) following published protocols

(Zhao et al. 2018). Parafilm was used as the membrane for blood feeding using defibrinated bovine blood in three treatments: Can-Alberta strain of EHDV-2, Florida strain of EHDV-2, and control (blood without the addition of virus). Infected blood was collected prior to feeding trials and from Hemotek feeders following each blood feeding session to monitor viral titer and to ensure titers were similar. The calculated titer of the two viral strain treatments were each 5.5 logs plaque forming units (PFU)/mL. Midges were allowed to feed on the Hemotek feeders for one hour prior to sorting from unfed midges using stereomicroscopy.

After feeding, cohorts of midges were anesthetized using a CO2 pad

(LabScientific Carbon Dioxide Staging Platform, Cat #BGSU-12, Highlands, NJ). Males and unfed females were discarded with the exception of 50 females maintained as unfed controls. Blood-engorged females were sorted into cups of ≤25 individuals separated by treatment. After blood-feeding, midges were provided 10% sucrose solution ad libitum and maintained at 25˚C with 12:12 light:dark cycle until processing.

Sample Collection

Samples collected from Culicoides included bodies, legs, and saliva to test for susceptibility to infection, disseminated infection, and transmission potential, respectively. Starting at day 3 post infection, cohorts of midges from each viral

110 treatment were selected at random daily for sample collection. On days 3 and 4, legs were dissected from bodies and both bodies and legs were placed into separate 1.5mL microcentrifuge tubes containing 150µL medium 199 and 5-10 2mm borosilicate glass beads (Sigma-Aldrich, St. Louis, MO) for homogenization. Beginning on day 5, capillary assays were performed to collect saliva. Upon removal of legs and wings, a capillary tube containing a small amount of immersion oil was applied to the mouthparts of each midge to collect saliva (Anderson et al. 2010). To induce salivation, a 0.01% solution of malathion in acetone was applied to the thorax of each individual (Boorman 1987).

Immersion oil containing saliva was blown out into 1.5mL microcentrifuge tubes containing 150µL of medium 199 (Wiggins et al. 2018).

Honey Cards

Honey cards were used alongside capillary assays to compare the two methods for saliva collection in Culicoides. Thus, the same individual midges were tested by both methods. Starting on day 5 post inoculation, midges in cups planned for collection the next day were transferred individually to small cages and given small pieces of filter paper (Whatman grade 1 filter paper, cut to <1cm2) coated in honey dyed blue with food coloring (McCormick & Co. Inc., Hunt Valley, MD) (Alto et al. 2017, Honorio et al. 2018).

Prior to dissection the following day, midges were inspected for blue abdomen coloration to indicate ingestion of blue honey. Honey cards of midges displaying blue coloration were collected into 1.5mL microcentrifuge tubes containing 150µL medium

199.

111

Sample processing

Body and leg samples were homogenized using the Bullet Blender Storm (Next

Advance, Troy, NY) for five minutes, following the manufacturers protocol. Honey cards were processed by crushing the filter paper with a pestle to release virus into solution.

Viral RNA was extracted using the Qiagen viral RNA mini kit (Cat#52906, Qiagen,

Hilden, Germany) following manufacturer’s protocols.

All samples were run by quantitative reverse-transcription polymerase chain reaction (qRT-PCR) using the SuperScript III One-step qRT-PCR kit (Thermo Fisher

Scientific, Waltham, MA). Reaction reagents included 2.2µL molecular grade water, 1µL

10µM forward primer, 1µL 10µM reverse primer, 10µL 2X reaction mix, 0.4µL EHDV probe, 0.4µL Platinum Taq/SuperScript III reverse transcriptase mix, and 5µL RNA template. Primers and probe sequences used were from Wernike et al. (2015) (F: AAA

AAG TTC YTC GTC GAC TGC, R: ATT GGC RTA RTA ACT GTT CAT GTT, Pr: ATC

GAG ATG GAR CGC TTY TTG AGA AAA T). Reaction conditions were modified from

Wilson et al. (2009) to 25 min at 55˚C, 2 min at 95˚C, and 45 cycles of 10 s at 95˚C and

1 min at 57˚C. The limitations for detection for the viral strains were determined at

Cq=35 for the Florida strain and Cq=37 for the Can-Alberta strain. Limitations of detection were determined through production of viral standards based on serially diluted virus samples. This approach allowed us to generate a standard curve to express the titer of EHDV in Culicoides samples by comparing cDNA synthesis for serial dilutions of EHDV in parallel with plaque assays to express viral titer as plaque forming unit equivalents (PFUe)/mL (Bustin 2000).

112

Statistical analysis

Contingency table analyses were run to test for strain, time, and time by strain interactions for bodies, legs, and saliva (honey cards and capillary assays positives combined and averaged). Legs were tested only if the body of the same individual already tested positive for EHDV viral RNA. As such, disseminated infection rates were calculated as the number of positive leg samples divided by the number of virus positive bodies detected. Saliva was tested only if both legs and bodies tested positive for viral

RNA and transmission rates were calculated by dividing total midges with detectable virus in saliva by total midges with disseminated infection detected in leg tissues.

Scatterplots were generated and linear regression analyses were run for each comparison to explore relationships between viral titers in bodies-legs and legs-saliva.

Welch’s t-tests with Welch-Satterthwaite degrees of freedom calculations were employed to look for significant differences in viral titers between viral strains in bodies, legs, and saliva. For t-tests, data were analyzed in two ways: (1) data aggregated by strain regardless of collection date and (2) dates aggregated by strain and early (day 3-

7) and late (day 8-13) periods. These periods were chosen based on an analysis of the progression of EHDV infections in C. sonorensis indicating that the number of virus particles reaches a maximum titer between 8-10 days post feeding (Mills et al. 2017).

Chi-square test of independence was employed to investigate whether transmission assay results were independent of strain. Results were found to be dependent on strain, so separate Chi-square analyses were used to determine whether there were differences in the number of positives identified through each transmission assay. Detected viral titer was compared between methods for Florida and Can-Alberta

113 strains using Welch’s two-sided t-tests. All statistics were run using R software (R Core

Team 2016, R 3.3.2 software, Vienna, Austria).

Results

Infection, Dissemination, and Transmission Rates

Culicoides infected with the Florida EHDV-2 strain had higher rates of infection than those infected with the Can-Alberta strain. Over the course of the experiment, 145 out of 167 (86.83%) bodies tested positive for the Can-Alberta strain of EHDV-2 and

237 out of 250 (94.80%) bodies tested positive for the Florida strain of the virus. By day three, when sampling began, 65.00% and 94.44% of bodies sampled were positive from

Can-Alberta and Florida strains, respectively. Infection rates varied by day (Table 5-1), reaching 100% of individuals for both strains on multiple sampling days. Contingency table analyses indicated significance in infection rates by strain (χ2=7.274, df=1,

P=0.007), time (χ2=20.125, df=11, P=0.044), and time by strain interactions (χ2=48.803, df=30, P=0.016). The bodies for the control midges (n=15) also were assayed and tested negative for both strains of the virus.

Dissemination of the two strains (leg samples) followed a pattern similar to infection in bodies, with higher dissemination observed in midges infected with the

Florida strain than the Can-Alberta strain. Dissemination to leg tissues was seen in 110 out of 145 individuals (75.86%) with the Can-Alberta strain and 208 out of 237 individuals (87.76%) with the Florida strain of EHDV-2. Dissemination for both the Can-

Alberta strain and Florida strain was recorded as early as day 3 post feeding, with

23.08% and 29.41% dissemination seen in the Can-Alberta and Florida strains, respectively. Also, by day five post feeding, infection and dissemination rates were

114 similar, suggesting a limited midgut escape barrier for these EHDV strains in C. sonorensis. Contingency table analyses indicated that dissemination rates were significant by strain (χ2=8.303, df=1, P=0.004), time (χ2=93.657, df=11, P<0.001), and time by strain interaction (χ2=105.15, df=30, P<0.001). Higher dissemination rates were seen for the Florida strain overall with dissemination surpassing 50% of individuals on day four for the Florida strain and day five for the Can-Alberta strain.

Transmission rates for the Florida strain were higher than those for the Can-

Alberta strain of EHDV-2 (combined assays). Positive saliva samples were detected for

47 out of 85 (55.29%) Can-Alberta infected midges and 92 out of 157 (58.60%) Florida strain infected midges. Transmission potential was not significant by strain overall

(χ2=0.130, df=1, P=0.719), but was significant by time (χ2=51.571, df=9, P<0.001) and time by strain interaction (χ2=64.932, df=24, P<0.001).

Viral Titer Analysis

Viral titers (PFU equivalents/mL) were significantly higher for Can-Alberta strain bodies (t=4.723, df=253.93, P<0.001), legs (t=5.355, df=177.9, P<0.001), and saliva

(t=8.599, df=78.755, P<0.001) overall, compared to Florida strain (Figure 5-1). When time periods were split into early infections and late infections, Can-Alberta strain body and saliva titers were both significantly greater than Florida strain (t=2.713, df=109.02,

P=0.008 and t=3.378, df=11.995, and P=0.005, respectively) in midges collected before day 8 and body, leg, and saliva titers were significantly higher after day 8 (body: t=4.063, df=144.08, P<0.001; leg: t=6.424, df=110.54, P<0.001; saliva: t=8.034, df=63.717, P<0.001). The only non-significant comparison was that of legs during the early time period (t=1.3164, df=67.915, P=0.1925).

115

Linear regression analyses indicated that viral titers in the bodies and legs of both the Can-Alberta and Florida strains were significantly positively correlated during the early period (Figure 5-2a and 4-2b, respectively) and late period (Figure 5-2c and 5-

2d, respectively). Low positive saliva sample sizes prohibited the analysis of leg by saliva regression analyses during the early period. Neither Can-Alberta (Figure 5-2e) nor Florida (Figure 5-2f) strains showed a significant correlation between leg and saliva titers during the late period.

Honey Card Versus Capillary Assay Comparison

For the 59 Can-Alberta strain and 121 Florida strain individuals with data for both assay methods, a significant interaction of strain and method was identified (χ2=13.459, df=3, P=0.004). When strains were analyzed separately, positive results were significantly dependent on method for both Florida (χ2=6.88, df=1, P=0.009) and Can-

Alberta strains (χ2=20.122, df=1, P<0.001). These results were confounding between strain with more positives detected by capillary assay in Can-Alberta strain infected midges (55.9% positive by capillary assay, 35.6% positive by honey card assay) and more positives detected by honey card assay in Florida strain infected midges (31.4% positive by capillary assay, 48.76% positive by honey card assay). Consensus in results for honey cards and capillary assays was seen in just over half (52.78%) of samples, with 33 individuals testing positive through both methods and 62 individuals testing negative through both methods. Of the remaining 47.22%, capillary assays detected positives in 38 samples and honey cards detected positives in 47 samples (Figure 5-3).

Significant differences were also detected in the titer of virus recovered using the two transmission assays. Mean honey card titer (Florida=2.604 log10 PFU/mL, Can-

116

Alberta=4.415 log10 PFU/mL) was significantly greater than capillary titer (Florida=1.620 log10 PFU/mL, Can-Alberta=3.212 log10 PFU/mL) for both the Florida (t=-6.47, df=80.781, P<0.001) and Can-Alberta (t=-4.735, df=38.082, P<0.001) strains. This trend was observed in saliva collections throughout the saliva collection period (7-13 days post feeding) except on day 7 when no Can-Alberta honey cards were positive (Figure

5-4).

Discussion

Although differences in the vector competence of Culicoides to multiple serotypes has been well documented (Paweska et al. 2005, Ruder et al. 2012, Ruder et al. 2015b), our data indicate that aspects of vector competence are also significantly affected by differences in viral strains within the same serotype. Infection and dissemination rates were significantly higher for the Florida strain of EHDV-2 than the reference Can-Alberta strain isolated in the 1960’s. Despite this, the Can-Alberta strain produced an overall higher titer of virus within the C. sonorensis at all stages of infection, including midgut, dissemination into body tissues, and saliva. It is unclear why an overall higher replicative advantage in viral titer in the Can-Alberta strain did not produce similar or higher infection, dissemination, and transmission rates than the

Florida strain of EHDV-2. Studies comparing two strains of West Nile virus (WNV) infection in mosquitoes observed similar patterns between rates of susceptibility to infection and association with viral titer (Veronesi et al. 2018). Specifically, WNV strain

FIN (Lim et al. 2013) infected Aedes japonicus Theobald at significantly lower rates than

WNV strain NY99 (Borisevich et al. 2006), despite the former strain producing higher titer infections. Along the same lines, higher viral titers of WNV strains in bodies and

117 legs were not associated with higher transmission rates (Richards et al. 2014). Taken together, these observations may suggest that different virus strains have different affinities to specific cells and organs within the insect and that viral titer is not always an accurate predictor of progression of infection.

Although C. sonorensis is a confirmed vector of EHDV broadly, the extremely high infection and dissemination rates seen for both strains of EHDV-2 were unexpected. The infection rates seen in similar studies investigating the same colony strain of C. sonorensis at 25˚C for EHDV-2 are variable (greatest daily percent positive ranged between 60% and 83% with inoculation titers of 106.9 PFU/mL and 106.8 TCID50, respectively) (Ruder et al. 2015b, Mills et al. 2017). Viral strain is one difference between previous and the present study that stands out as a contributing factor to the variation in infection rates. This underscores the importance of considering the origin of an EHDV viral strain when designing and conducting vector competence experiments.

The high infection, dissemination, and transmission rates seen in the present study also have implications for understanding barriers to transmission to EHDV within this species. Culicoides sonorensis, which is also a documented bluetongue virus vector

(Bowne and Jones 1966), was shown to have a midgut infection and midgut escape barrier and lack a salivary gland infection and escape barrier to bluetongue virus (Fu et al. 1999). The overall high infection and dissemination rates seen at 3 days post feeding and thereafter indicate a very low midgut infection and escape barrier in C. sonorensis to both strains of EHDV used in this study. Transmission rates were moderately high

(55.29% of Can-Alberta infected midges and 58.60% of Florida infected midges).

Although these overall transmission rates are considerably lower than the overall

118 infection and disseminated infection rates, transmission rates increased considerably for both strains at ten days post feeding indicating that a salivary gland barrier is unlikely for either strain of EHDV-2 in C. sonorensis.

While rates of infection and dissemination were overall higher for the Florida strain, greater viral titers were recovered from individuals infected with the Can-Alberta strain of the virus. This finding is an important consideration for overall transmission potential between the two strains. The average viral titer of EHDV-2 in the saliva of

Florida-strain infected Culicoides was 2.21 log PFU equivalents/mL (3.16 log10 TCID50), which is above the accepted transmission threshold of 2.7 log10 TCID50 (Jennings and

Mellor 1987, Ruder et al. 2012), but much lower than the viral titer in Can-Alberta infected midges of 3.59 log PFU equivalents/mL (5.13 log10TCID50). One study investigating the transmission of bluetongue virus to sheep by C. sonorensis found that greater viral titers found in infected midges resulted in greater antibody responses and higher titer viremia in sheep (Baylis et al. 2008). While both viral strains surpass the viral threshold for transmission, the overall higher titer of virus present in the saliva of

Can-Alberta infected midges may make this strain more virulent to EHDV hosts.

Culicoides sonorensis is not found in great abundance in the state of Florida (Smith and

Stallknecht 1996, Smith et al. 1996b, Borkent and Grogan 2009, Vigil et al. 2014, Sloyer et al. 2018), indicating that the infection and transmission dynamics may be affected by coevolution of the Florida strain with other vectors. Investigating the Florida strain transmission rate and titer of abundant Floridian species that have coevolved with this viral strain may provide greater insight into the transmission dynamics observed in this study.

119

Honey FTA cards have successfully been employed with C. sonorensis for detection of Schmallenberg virus transmission (Veronesi et al. 2013); however, virus preserving agents present in FTA cards have been linked to premature death in mosquitoes indicating that this substrate may not be suitable for arthropod transmission studies (Alto et al. 2017). The present study showed that virus can be collected and detected using plain filter papers without the need for specialized preserving agents, which has implications for future vector competence studies. Capillary assays rely on the survival of the vector throughout the study period as this is an end stage assay that can only be performed on a living individual immediately preceding death (Colton et al.

2005, Alto et al. 2018). Due to challenges with maintaining wild collected Culicoides in laboratory environments for extended periods of time (Tanya et al. 1992), this assay may limit transmission sample sizes. Honey cards are not a terminal assay and can be employed throughout an infection study to monitor for first transmission events without the need to harvest the individual (Brustolin et al. 2017, Guedes et al. 2017, Honorio et al. 2018). This non-destructive sampling technique enables investigators to test the same individual multiple times and to calculate the extrinsic incubation period on a per capita basis. Further, the honey card approach is much less labor intensive than performing capillary tube assays and allows for high throughput of tests of saliva.

Finally, as this assay is not terminal, use of this method could lead to greater transmission sample sizes in future studies on the vector competence of Culicoides when wild-collected individuals are being used.

Ultimately, the honey card method was only significantly better for Florida infected midges in the present study. It is unclear why confounding results were found in

120 transmission testing methods between strains; however, for both strains, a large proportion of transmission events were only detected by a single method. Based on the confounding results of the comparison and the inability of either method to detect every positive result, using multiple methods to detect transmission when possible may provide the most accurate representation of the transmission rate (Honorio et al. 2018,

Wiggins et al. 2018).

The present study demonstrated the susceptibility of C. sonorensis to infection by an actively circulating strain of EHDV isolated in the Florida panhandle. Furthermore, when compared with a reference strain of EHDV from Alberta, Canada, rates of infection and dissemination were significantly higher for the Florida strain of the pathogen while viral titers were higher for the Can-Alberta strain. These results demonstrate that strain plays a significant role in the overall vector competence in

Culicoides and should be considered when developing future transmission studies.

121

Table 5-1. Percent daily infection (body), dissemination (legs), and transmission (saliva) rates (combined capillary assay and honey card detected transmission) for the Can-Alberta strain and the Florida strain of EHDV-2 in Culicoides sonorensis. Can-Alberta Strain Florida Strain Day # Body Legs Salivaa # Body Legs Salivaa 3 11 65.00 23.08 nd 18 94.44 29.41 nd 4 19 80.00 41.67 nd 18 94.44 76.47 nd 5 12 91.67 90.91 22.22 24 91.12 86.36 25.00 6 12 91.67 100.00 18.18 17 94.12 87.50 0.00 7 13 100.00 85.71 58.33 18 100.00 94.44 35.29 8 18 76.92 70.00 50.00 19 100.00 100.00 52.63 9 23 100.00 84.21 66.67 18 100.00 88.89 53.33 10 23 84.21 87.50 75.00 23 91.30 90.48 78.95 11 15 100.00 90.00 87.5 24 87.50 95.24 75.00 12 11 100.00 81.82 87.5 27 88.89 100.00 70.59 13 9 88.89 87.50 50.00 28 100.00 91.30 75.00 14b nd nd nd 16 100.00 100.00 81.25 nd=not determined aTransmission assays were not started until day five post feeding bNo Can-Alberta infected individuals survived to day 14 post feeding

122

Figure 5-1. Viral titers in the bodies, legs, and saliva of Culicoides sonorensis infected with the Can-Alberta or Florida (FL) strain of EHDV-2. Titers overall, during the first 3-7 days post feeding (early), or during the last 8-13 days post feeding (late) are shown for both strains. Error bars indicate the standard error of each mean.

123

Figure 5-2. Viral titer relationships between bodies, legs, and saliva of Culicoides sonorensis infected with two strains of EHDV-2. a) Body by leg titer for Can- Alberta, b) Body by leg titer for Florida (FL) strain, c) Body by leg titer for Can- Alberta, d) Body by leg titer for FL, e) Leg by saliva for Can-Alberta, f) Leg by saliva for FL. Figure 5-2a-b represent the early period (day 3-7 post feeding) while figure 5-2c-f represent the late period (day 8-13 post feeding).

124

Figure 5-3. Virus detection by saliva collection method for two strains of EHDV-2 in Culicoides sonorensis. Combined displays all data from both strains. Consensus between the results of capillary assays and honey cards was reached for 53% of samples on average.

125

Figure 5-4. Honey card and capillary assay titers for two EHDV strains in Culicoides sonorensis over time. Samples were collected daily (days 7-13 post feeding). Culicoides sonorensis individuals were initially fed 5.5 log PFU/mL of each strain (Can-Alberta and Florida (FL) strains of EHDV-2). Error bars show the standard deviation of the mean.

126

CHAPTER 6 VECTOR COMPETENCE OF CULICOIDES INSIGNIS (DIPTERA: CERATOPOGONIDAE) FOR EPIZOOTIC HEMORRHAGIC DISEASE VIRUS SEROTYPE 2

Epizootic hemorrhagic disease virus (EHDV) is an arbovirus in the family

Reoviridae, genus Orbivirus that primarily affects ungulates of the families Cervidae and

Bovidae (Savini et al. 2011). Culicoides biting midges are the established vector genus for this pathogen, although patchy distribution of known vector species indicates that alternative Culicoides vectors are likely present in North America (Pfannenstiel et al.

2015). Identifying alternative EHDV vector species is an important step towards developing sound management strategies tailored for specific biological traits of the implicated vector.

The only confirmed vector of EHDV in North America is Culicoides sonorensis

Wirth & Jones (Foster et al. 1977, Jones et al. 1977). While C. sonorensis has been identified as a competent vector of multiple EHDV serotypes (Ruder et al. 2012; 2015b;

2016), this species is primarily established west of the Mississippi River (Borkent and

Grogan 2009). Although scattered populations do occur east of the Mississippi River

(Borkent and Grogan 2009, Vigil et al. 2014), C. sonorensis has been rare or absent during multiple large scale Culicoides sampling studies in the southeastern United

States (Smith and Stallknecht 1996, Smith et al. 1996b, McGregor et al. 2018a, Sloyer et al. 2018). Alternative vector species are likely to occur in this region, necessitating investigations into the ability of resident species to become infected with and transmit this virus.

While the southeastern United States has a broad diversity of Culicoides species

(Blanton and Wirth 1979, Borkent and Grogan 2009, Grogan Jr. et al. 2010), a subset of

127 species stands out as vector candidates due to their preferential feeding on EHDV hosts and their abundance during periods of active EHDV transmission. These species include Culicoides biguttatus (Coquillett), Culicoides debilipalpis Lutz, Culicoides obsoletus Meigen, Culicoides pallidicornis Kieffer, Culicoides paraensis Goeldi,

Culicoides spinosus Root & Hoffman, Culicoides stellifer (Coquillett), and Culicoides venustus Hoffman (Jones et al. 1983, Mullen et al. 1985a, Smith and Stallknecht 1996,

Smith et al. 1996b, Pfannenstiel et al. 2015, McGregor et al. 2018b). Limited infection studies have been conducted on some of these species. One laboratory infection study was carried out on C. debilipalpis (as C. lahillei) for EHDV-2 and found that at high titers of virus, this species is a weakly competent vector of EHDV (Smith et al. 1996a). A New

York population of Culicoides venustus was tested for competence for EHDV and bluetongue virus (BTV) in the laboratory, with only one individual becoming infected with

EHDV-1 (Jones et al. 1983). Transmission potential was not assayed for either species.

While these species have been implicated previously, they are not the only species present in great abundance near EHDV hosts and additional potential vector species should be identified.

Another species that has been found in close and abundant association with

EHDV hosts such as cattle and white-tailed deer is Culicoides insignis Lutz (Kramer et al. 1985a, Greiner et al. 1990, Sloyer et al. 2018). This species occurs in the far southeastern states of Alabama, Florida, and Georgia (Borkent and Grogan 2009), states that have historically experienced recurrent outbreaks of EHDV (Prestwood et al.

1974, Smith and Stallknecht 1996), and evidence indicates recent northwestern range expansion into Mississippi and Louisiana (Vigil et al. 2018). Culicoides insignis is also a

128 confirmed vector of BTV (Tanya et al. 1992). BTV and EHDV are closely related orbiviruses, which affect wild and domestic ruminants, that possess proteins showing greatly conserved sequences both within and between virus groups (Wilson 1991). The close antigenic relationship between these two viruses led to the early postulation that

Culicoides, the known vector genus for BTV, could also be the vector group for EHDV

(Moore and Lee 1972). Furthermore, Culicoides species that are competent for one virus are often competent for the other as in Culicoides bolitinos Meiswinkel, Culicoides brevitarsis Kieffer, Culicoides imicola Kieffer, and C. sonorensis (Foster et al. 1968,

Jones et al. 1977, Parsonson and Snowdon 1985, Muller 1987, Paweska et al. 2005,

Venter et al. 2006). For these reasons, this species is a candidate EHDV vector and warrants investigation.

The goal of this experiment was to determine the vector potential of C. insignis for EHDV serotype 2 by feeding wild-collected individuals viremic blood and testing for infection, dissemination, and transmission potential. Prior to completing the infection study, a series of blood-feeding trials were carried out to determine the optimal membrane, blood type, and starvation conditions to facilitate successful laboratory blood feeding for C. insignis. Various blood feeding methods have been tested for feeding field-collected Culicoides with variable success including Hemotek systems

(Pagés et al. 2018), blood-soaked cotton pledgets (Venter et al. 2005), and novel blood- feeding apparatuses (Venter et al. 1991). In order to successfully complete a vector- competence experiment, an effective method of blood feeding the putative vectors must be determined.

129

Methods

Blood Feeding Trial Specimen Collection

Wild Culicoides were collected from two sites in the Florida peninsula that historically have produced large numbers of Culicoides insignis: (1) MacArthur Agro-

Ecology Research Center near Okeechobee, Florida and (2) a deer farm located in

Marion County, Florida. Culicoides were collected using Centers for Disease Control and Prevention (CDC) miniature light traps (Model 2836BQ, BioQuip Inc., Rancho

Dominguez, CA) with LED black light arrays (Model 2790V390, BioQuip Inc.) and baited with CO2 (solid dry ice). CDC traps were modified with mesh to exclude large arthropods and collected insects directly into a BugDorm insect cage (Model #4F2222,

BugDorm, Taiwan) (Erram and Burkett-Cadena 2018). Live-captured Culicoides were aspirated into paper cups with no-see-um netting and provided with 10% sucrose.

Blood Feeding Apparatus Design

A blood-feeding apparatus was designed following the protocols of Venter et al.

(1991) and was modified using readily available laboratory materials. Blood-feeders were constructed from 15mL conical tubes cut off at the 6mL mark and modified with mesh-covered air holes along the sides and a latex port for aspirating midges into the feeder. The blood-feeders were suspended in the blood meal using a stand constructed from an aspirator collection cup with the bottom removed. The whole feeding apparatus was placed inside a plastic vessel containing blood and a magnetic stir bar for promoting circulation of the blood meal. The vessel containing the blood was placed inside a secondary container holding water, which was placed onto a hot plate (Figure

6-1). A thermometer was inserted into the blood at all times and the hot plate was

130 adjusted accordingly to maintain the blood at approximately 38˚C, the average body temperature for white-tailed deer (Wilber and Robinson 1958).

Blood Feeding Trial Variables

The variables tested included membrane, blood, and starvation length.

Membranes included hog casings (The SausageMaker Inc, Buffalo, NY), lamb-derived condoms (Trojan, Ewing, NJ), Parafilm (Bemis Company Inc, Oshkosh, WI), nitrile gloves (Fisher Scientific, Pittsburgh, PA), 1-day old chick skin (Biochemed Services,

Winchester, VA), mouse skin (Biochemed Services), and cow skin (Biochemed

Services). Blood types tested included defibrinated cow, defibrinated horse, and defibrinated sheep (Hemostat Laboratories, Dixon, CA). Starvation of midges (no blood or sugar solution provided) at 6, 12, 24, and 36 hours prior to blood-feeding were tested.

Midges were allowed to blood feed for a period of one hour in dark or semi-dark conditions (blood feeding apparatus in a lighted room, covered with a box to provide darkness) prior to knockdown with triethylamine (TEA). Blood engorged individuals were then counted.

Infection Trial Culicoides insignis Collections

Culicoides insignis individuals were collected using the same methods as in the blood-feeding trials described previously. Midges were aspirated into 475mL paper cups and held at 25˚C and 60-80% RH prior to blood-feeding. Upon collection, midges were fed water only until blood-feeding, which occurred within 4 hours of trap retrieval

(generally 10-11:00 am).

131

Viral Blood Feeding

Epizootic hemorrhagic disease virus serotype 2 was isolated by the Cervidae

Health Research Initiative from the spleen of a white-tailed deer that died of the disease in Gadsden County, Florida, in 2016. Plaque assays were run to produce viral standards for titer determination. The virus was passaged twice on African green monkey kidney (Vero) cells before being used for viral blood feeding using defibrinated bovine blood. Pre- and post-blood feeding samples were collected from the feeding apparatus to determine the titer of virus given to the midges during each trial.

Midges were left to feed in the feeding apparatus (Figure 6-1) for 1.5 hours in dark conditions inside an incubator set at 25˚C and 60-80% humidity. At the end of the feeding period, midges were anesthetized using triethylamine (TEA) and blood engorged midges were separated and identified to species using morphological characteristics (Blanton and Wirth 1979) before being placed in individual 120mL paper cups with no-see-um netting.

Sample Processing

Midges were processed 10 days after viral blood feeding or within 12 hours of their death if they died prior to the full 10-day incubation period. Midge bodies were kept to test for initial infection and legs were separated to test for disseminated infection. For each individual, the body and all six legs were collected into 1.5mL microcentrifuge tubes (one tube for the body and one tube for all legs combined) containing 150µL medium 199 (HyClone Medium 199, GE Healthcare Life Sciences, Logan, UT) and five to ten 2mm borosilicate glass beads (Sigma-Aldrich, St. Louis, MO) for homogenization.

132

Transmission was assayed in two ways: (1) using honey cards administered daily to monitor for transmission events and (2) using capillary assays at harvesting, when possible. Both methods were used to increase the chance of detecting virus in the saliva of infected midges.

Honey cards were placed on each cup containing an individual midge from day three until either natural death occurred or until the incubation period ended. Honey cards were made from small pieces of filter paper (Whatman grade 1 filter paper, cut to

<1cm2) coated in honey dyed blue with food coloring (McCormick & Co. Inc., Hunt

Valley, MD). Each honey card was collected into a 1.5mL microcentrifuge tube containing 150µL medium 199.

Capillary tubes were administered following the protocols of Anderson et al.

(2010). In brief, after removal of legs and wings, anesthetized midges were moved to a

Plexiglas board and placed on a piece of tape to prevent movement. A capillary tube containing immersion oil was applied to the mouthparts of the midge to collect saliva.

Midges were left to salivate for one hour prior to removal of the midge body and saliva into separate 1.5mL microcentrifuge tubes containing 150µL medium 199. Immersion oil containing saliva was blown out of capillary tubes using a p200 pipette (Gilson

Pipetman, Middleton, WI).

RNA Extraction and Virus Detection

Body and leg samples were homogenized separately using the Bullet Blender

Storm (Next Advance, Troy, NY) for five minutes following manufacturer’s protocols, while honey cards were processed by crushing the honey card in the media using a pestle to release virus into solution. All samples underwent RNA extraction using the

133

QiaAmp Viral RNA Mini Kit (Cat#52906, Qiagen, Hilden, Germany) following manufacturer’s protocols.

Quantitative reverse-transcription polymerase chain reaction (qRT-PCR) was run on all samples using the SuperScript III One-step qRT-PCR kit (Thermo Fisher

Scientific) following established protocols (described in Chapters 3 and 4). Primer and probe sequences were adapted from Wernike et al. (2015) and reaction conditions were modified from Wilson et al. (2009) to 25 mins at 55˚C, 2 mins at 95˚C, and 45 cycles of

10 s at 95˚C and 1 min at 57˚C.

Statistical Analyses

For the blood-feeding trials, statistical analyses were only conducted to investigate the effect of membrane and starvation period on feeding success. Blood type was not analyzed as horse and sheep blood were not used often enough for valid comparisons. Data were determined to not follow a normal distribution, so Kruskal

Wallis tests followed by Wilcoxon rank sum tests were conducted to investigate significance in membrane types.

For the C. insignis infection experiments, four separate trials were conducted on field-collected C. insignis using different viral titers and two different midge populations.

Infection, dissemination, and transmission rates were calculated for each trial. Chi- square tests of independence were used to investigate whether infection, dissemination, and transmission rates were significantly different between populations and inoculation titers. All statistics were run using R software (R Core Team 2016, R 3.3.2 software,

Vienna, Austria).

134

Intrathoracic Inoculation

A subset of midges was subjected to intrathoracic inoculation in order to increase sample sizes for the investigation into transmission potential. Midges were anesthetized using a CO2 pad and injected with 45nL of viral media (5.0 log PFU equivalents/mL) using the Nanoject II auto-nanoliter injector (Drummond Scientific, Broomall, PA).

Needles were pulled from 9cm long glass capillary tubes (Drummond Scientific,

Broomall, PA) and were changed after every five individuals to prevent the needles from becoming too blunt and causing increased injury to midges. Injected midges were placed into individual 120mL paper cups with no-see-um netting and monitored for 24 hours to detect injection-related mortality. Those individuals that survived the initial 24- hour period were provided honey cards at 2 days post inoculation. Fresh honey cards were provided daily to monitor for transmission events until the death of the midge.

Whole midge bodies and honey cards were processed as described above and viral titers were determined based on viral standards.

Virus Screening of Field-Collected C. insignis

Pool screening of field-collected C. insignis for EHDV was used to determine whether virus was present in the wild populations from which experimental specimens were taken. Midges were sorted into pools of ≤50 individuals and placed into 500µL medium 199 with Earle’s Balanced Salt Solution with ten to twenty 2mm borosilicate beads for homogenization. Protocols for RNA extraction and qPCR were identical to those described above for experimentally infected midges.

135

Results

Blood Feeding Trials

A combined total of 48 trials were run with 2,307 C. insignis females (Table 6-1).

Due to inconsistency in the number of midges collected per trial, not all combinations were tested. All membranes were tested with cow blood, four membranes with horse blood, and two with sheep blood. Overall, the combination with the greatest feeding success was cow blood with a 1-day old chick skin membrane with an 18.75% feeding success rate. The poorest feeding success was achieved with the combination of mouse skin membrane with either cow or sheep blood. No midges fed on this membrane with either blood type (N=19 with cow blood, N=24 with sheep blood).

Overall, a significant association was found between blood-feeding rates and membrane type (Kruskal-Wallis χ2=17.964, df=8, P=0.022). In pairwise comparisons, the sheep-derived condom performed significantly better than cow skin (P=0.009), nitrile

(P=0.009), and Parafilm (P=0.020) and the chick skin membrane was significantly better than the Parafilm membrane (P=0.046). No other pairwise comparisons were significant at α=0.05 (Table 6-2). No significant association was found between blood-feeding rates from midges starved for 0, 6, 24, and 36 hours (Table 6-3, Kruskal-Wallis χ2=5.969, df=3, P=0.113).

Infection, Dissemination, and Transmission

Four trials were run to achieve adequate sample sizes for analysis (Table 6-4). A total of 18 individuals from the Okeechobee population blood fed during trial one. The

EHDV viral titer for this trial was determined to be 5.05 log PFU equivalents/mL. The infection rate for this first trial was high with 17/18 (94.44%) individuals showing virus

136 positive bodies. Dissemination rates for trial one were also high at 11/17 (64.71%) infected individuals showing a disseminated infection. The transmission rate for the first trial was 5/11 individuals (45.45%). Mean viral titer for bodies was determined to be

2.00 PFU/mL, legs were 2.14 PFU/mL, and 1.35 PFU/mL for saliva (Figure 6-2).

During the second trial, 64 midges from the Okeechobee population fed on blood with a calculated viral titer of 3.03 log PFU equivalents/mL. Of the 64 blood-engorged midges, only 14 had detectable infections in the body tissues (21.88%). Out of these 14 individuals, only one had a disseminated infection (7.14%) and the same individual had virus-positive saliva. High mortality occurred during the second trial that resulted in 50 individuals (78.13%) dying by the 4th day post feeding. For this trial, mean body titer was determined to be 2.04 PFU/mL. Only one individual had positive legs and saliva for this trial, at 1.72 and 1.14 PFU/mL, respectively (Figure 6-2).

The blood for trial three had a titer of 2.94 log PFU equivalents/mL with a sample size of 62 blood engorged individuals from the Ocala population. Five of the 62 had

EHDV positive bodies (8.07%). One of those five infected individuals had a disseminated infection in the legs (20%). The same individual had detectable virus in the saliva identified through honey cards. The mean body titer was determined at 1.87 log PFU/mL. The single positive leg and saliva sample were 1.43 log PFU/mL and 1.17 log PFU/mL, respectively (Figure 6-2).

A total of 93 individuals from the Ocala population blood fed during trial four with a determined blood titer of 4.00 log PFU/mL. Four of the 93 individuals (4.3%) had infected bodies at harvest and of those, two (50%) had disseminated infections in the leg tissue. Neither individual had detectable virus in the saliva through honey cards or

137 capillary tube assays. Mean body titer was calculated as 1.63 log PFU/mL and mean leg titer was 1.49 log PFU/mL (Figure 6-2).

Capillary tube assays were only administered to individuals who survived throughout the entire experimental period. In total, 19 capillary tube assays were performed on midges at 10 days post inoculation. No positive results were recovered from saliva collected by capillary assay.

Significant differences were identified between the Ocala and Okeechobee populations at low (χ2=29.559, df=2, P<0.001) and high virus titer (χ2=58.331, df=2,

P<0.001). Infection was significantly higher in the Okeechobee population at low titer

(χ2=6.368, df=1, P=0.012). Dissemination and transmission rates were not analyzed as the high mortality that occurred on day 4 post inoculation likely affected the detectable dissemination and transmission rates. At high titer, significantly greater rates of infection

(χ2=82.289, df=1, P<0.001), dissemination (χ2=4.413, df=1, P=0.036), and transmission

(χ2=45.45, df=1, P<0.001) were detected in the Okeechobee population, although it is important to note that the high titer given to the Okeechobee population was 5.05 log

PFU/mL compared with the high titer given to Ocala of 4.00 log PFU/mL.

Due to significant differences between populations, analyses between different inoculation titers were conducted within populations. Significant differences in infection rates were detected in the Okeechobee population between individuals fed 3.0 log

PFU/mL and 5.0 log PFU/mL blood meals (χ2=45.263, df=1, P<0.001). Dissemination and transmission were not analyzed, again due to the high mortality that occurred on day 4 post inoculation. Within the Ocala population, there was no significant difference

138 in infection rates (χ2=1.149, df=1, P=0.284) between midges fed 3.0 log PFU/mL and

4.0 log PFU/mL.

Microinjection

A total of 28 individuals were microinjected, with 12 individuals surviving until at least day 2 post inoculation to monitor saliva for transmission through honey card assays. Five of the 12 surviving individuals (41.67%) had detectable virus in the saliva.

Two of the five individuals had detectable virus at two days post inoculation (DPI), the other three individuals had detectable virus at 3 DPI.

Virus Screening of Field-Collected Midges

A total of 17,800 wild C. insignis sorted into 356 pools of 50 individuals were tested for the presence of EHDV viral RNA using qRT-PCR. There were no positive samples detected from this effort indicating that it is very unlikely that any of the experimental midges were infected with EHDV prior to assays.

Discussion

The sheep-intestine derived condoms resulted in high blood-feeding rates and provided significantly higher infection rates than multiple other membranes. This membrane also has been used with success in other Culicoides studies (Hunt and

McKinnon 1990). The ease of availability in commercial markets makes this membrane well suited for experimental use. It is important to note that when using this membrane, thoroughly washing and drying the membrane prior to use is necessary as midges will stick to moist membranes.

139

One-day old chick skins were also a good membrane for successful blood- feeding of Culicoides insignis. This membrane also has been found to work effectively with other Culicoides species (Hunt and McKinnon 1990, Venter et al. 1991). However, there are drawbacks to this membrane, such as availability, that limits its utility.

Decontaminating chick skins while also maintaining Culicoides feeding efficacy is challenging and often not possible, limiting each membrane to a single use. Finally, there are ethical problems associated with the use of animal products, especially those situations where an animal is explicitly harvested for scientific purposes.

Although the hog casing membrane resulted in moderate feeding success with C. insignis, this membrane is not recommended for feeding Culicoides. Despite attempts to dry the membrane prior to use, this membrane repeatedly became too moist during blood-feeding resulting in high mortality. Hog casings have been used as a membrane for numerous studies conducted on mosquitoes (Mahmood et al. 2004, Alto et al. 2017,

Wiggins et al. 2018) and is relatively inexpensive and easy to work with. However, based on the moisture-induced mortality seen in this study, its use should be limited to larger insects.

The other membranes tested performed poorly overall. The cow skin membrane resulted in extremely low blood-feeding success. Culicoides insignis is known for its association with cattle (Kramer et al. 1985a, Saenz and Greiner 1994), an observation that led to the use of this membrane in blood-feeding trials. Efforts were made to remove underlying fat and tissue to provide a thinner membrane for blood-feeding prior to trials. Further attempts to provide a naturally thinner mammalian skin membrane by using mouse skin also were unsuccessful. Both artificial membranes performed poorly

140 with field-collected individuals. Colonized Culicoides sonorensis are frequently fed using

Parafilm membranes (Fu et al. 1999, Ruder et al. 2015b, Mills et al 2017) indicating that the mouthparts of Culicoides are capable of cutting through this membrane. This lack of feeding through artificial membranes by field-collected specimens may be linked to the nonporous nature of these materials.

Starvation period did not have a significant association with any of the blood- feeding rates, which has important implications for the use of field-collected midges in future studies. Culicoides insignis survival in a laboratory environment has been documented to be fairly low in previous vector competence experiments (Tanya et al.

1992). Starvation may cause stress to field-collected midges that may shorten the lifespan or lead to fewer individuals surviving until blood-feeding leading to lower sample sizes overall.

Ultimately, the results of the blood feeding trials indicated that the best membranes for feeding C. insignis individuals using the feeder designed by Venter et al.

(1991) were the 1-day old chicken skin membrane and the lamb-derived condom membrane. Both membranes resulted in relatively high feeding rates. While the hog casing membrane resulted in moderate feeding rates, many midges were lost to the moisture that accumulated on this membrane indicating that this membrane has low utility for Culicoides blood feeding.

This is the first study to investigate the vector competence of C. insignis for

EHDV serotype 2. Culicoides insignis is a confirmed vector of bluetongue virus (Tanya et al. 1992), a closely related virus in the genus Orbivirus. Based on the results of our

141 study, C. insignis is also a weakly competent vector of EHDV in the laboratory, supporting replication of the pathogen in the body, legs, and saliva.

The disparity in infection rates between different viral titers has implications for the likelihood of C. insignis as a vector of EHDV. At high titers (5 log PFU/mL titer in the first trial), C. insignis had high rates of infection (94%), dissemination (73%), and transmission (45%) indicating little to no active barriers to infection or transmission.

However, at lower titers (3-4 log PFU/mL), lower infection rates (4-22%) were detected overall indicating the likelihood of dose-dependent barriers to transmission, a phenomenon seen in other vector-borne pathogen systems (Hardy et al. 1983). Midgut infection barriers have been documented for other Culicoides-borne pathogens, and based on the results of this study, likely also exist in C. insignis for EHDV (Fu et al.

1999).

It is challenging to infer whether salivary gland barriers exist in the Ocala population or if low transmission rates were simply a consequence of low infection and dissemination rates for this population. In trial 4 (4 log PFU/mL trial with the Ocala population), no transmission was detected in study midges. Very low transmission (one individual) was detected at 3 log PFU/mL in the Ocala population. Previous studies have found no evidence of the presence of salivary gland infection and escape barriers in C. sonorensis for BTV (Fu et al. 1999). No Ocala midges were microinjected for this study. However, Okeechobee C. insignis fed 5.0 log PFU/mL and microinjected

Okeechobee midges had transmission rates of 41-45% indicating that even at high inoculation rates, salivary gland barriers reduce transmission of EHDV-2 in C. insignis.

142

Significantly higher infection rates were detected in the Okeechobee population overall, indicating that this population may be more susceptible to infection with EHDV-

2. Unfortunately, due to the mortality event that occurred on the fourth day of trial two, it is unclear whether comparative dissemination and transmission rates followed a similar trend at 3 log PFU/mL. Variation in vector competence among different populations has been shown in a variety of vector-borne pathogen systems, including Culicoides-borne pathogens (Jones and Foster 1978, Venter et al. 2009a, Alto et al. 2017). The

Okeechobee and Ocala populations are separated by around 250km indicating that interbreeding between these two populations is extremely unlikely and that environmental pressures are acting on, and affecting, these populations differently.

While C. insignis seems to be a somewhat competent vector of EHDV at high titers, virus was not recovered from any field-collected pools. There were no documented EHDV related deer deaths during the time of midge collections, indicating that if virus were present in the midge population it would be at extremely low levels.

Additionally, this species is extremely abundant throughout most of its range in Florida

(Kramer et al. 1985b, Garvin and Greiner 2003, Sloyer et al. 2018). Therefore, unless large scale outbreaks are occurring, the likelihood of recovering infected individuals would be low.

Although drawing inferences on viral titers in saliva for trials 2, 3, and 4 is challenging due to low sample size, the mean detected viral titer for the first trial (5

PFU/mL in Okeechobee midges) was only 1.35 log PFU equivalents/mL (1.93 log10

TCID50; n=5 individuals). This viral titer is considered too low on average to affect transmission of EHDV to a competent host, as the minimum saliva titer necessary for

143 transmission is believed to be 2.7 log10 TCID50 (Jennings and Mellor 1987, Ruder et al.

2012). Two of the 237 individuals assayed did produce sufficiently high salivary titers to facilitate transmission (4.11 log10 TCID50 and 2.93 log10TCID50) indicating that some C. insignis individuals would be capable of transmitting infections to susceptible hosts in nature. Due to low survival of field-collected midges, each trial was only carried out for ten days, which has been found to provide adequate time for a transmissible infection with BTV infected Culicoides (Foster et al. 1963). However, evidence from BTV studies indicates that 14 days extrinsic incubation period may be necessary for successful transmission (Foster and Jones 1973, Tanya et al. 1992). The shortened incubation period in the present study may have artificially lowered the detectable virus that would be present if longer incubation were allowed.

Based on the results of this study, C. insignis is a weakly competent vector for

EHDV. Some individuals at all titers became infected with and maintained a disseminated infection to EHDV-2 in the laboratory. However, transmission was low in trials at 3-4 log PFU equivalents/mL. It has been speculated that multiple species could be involved in the transmission of BTV in southern Florida (Tanya et al. 1992). Based on low infection, dissemination, and transmission rates at low inoculation titers and low saliva titers even when given high titer inoculation, other vectors of EHDV are likely present in southern Florida. Additional testing on other Culicoides species throughout the state of Florida is warranted to better understand the epizootiology of EHDV in this state.

144

Figure 6-1. Blood-feeding apparatus used for blood-feeding trials and Culicoides insignis vector competence trials. 1. Blood feeding tube with hole for aspirating midges into feeder at the top, mesh covered air holes on the side, and membrane at bottom; 2. Full blood feeding apparatus deconstructed including A) water bath, B) blood meal container, C) thermometer, D) magnetic stir bar, E) feeding tube stand, and F) blood feeding tubes with cotton and tape applied to aspirator holes at top; 3. Full blood feeding apparatus constructed on hot plate.

Table 6-1. Culicoides insignis blood feeding trial of membrane and blood combinations, showing percentage fed with total fed over total attempted in parenthesis. Combinations not tested are denoted as nt. Chick skins were taken from 1-day old animals and the condoms used were lamb-derived. Membrane Cow Blood Horse Blood Sheep Blood Total

Chick Skin 18.75% (18/96) nt 10.00% (2/20) 17.24% (20/116) Condom 4.45% (38/854) 6.89% (2/29) nt 4.53% (40/883) Cow Skin 0.68% (1/148) 0.00% (0/17) nt 0.61% (1/165) Hog Casing 6.96% (27/388) 0.00% (0/6) nt 6.85% (27/394) Mouse Skin 0.00% (0/19) nt 0.00% (0/24) 0.00% (0/43) Nitrile Glove 0.28% (1/359) nt nt 0.28% (1/359) Parafilm 1.35% (4/297) 0.00% (0/16) nt 1.28% (4/313) Total 4.12% (89/2,161) 2.94% (2/68) 4.55% (2/44) 4.09% (93/2,273)

145

Table 6-2. Pairwise Wilcoxon rank sum tests on blood-feeding rates of Culicoides insignis by membrane type. Chick skins used for this assay were from 1-day old animals and the condoms used were lamb-derived. Chick Skin Condom Cow Skin Hog Casing Mouse Skin Nitrile Condom 0.324 - - - - - Cow Skin 0.073 0.009 - - - - Hog Casing 0.145 0.118 0.347 - - - Mouse Skin 0.160 0.050 0.724 0.347 - - Nitrile 0.073 0.009 1.000 0.273 0.724 - Parafilm 0.046 0.020 0.810 0.382 0.548 0.810

Table 6-3. Culicoides insignis blood feeding trial of starvation periods. Effect of starvation on blood feeding success was only measured using cow blood. Chick skin membranes were taken from 1-day old animals and condoms were lamb-derived. Membrane 0 hours 6 hours 24 hours 36 hours Total Chick Skin 18/89 2/27 0 0 20/116 Collagen 0 4/34 0 0 4/34 Condom 27/686 7/78 4/84 2/35 40/883 Cow Skin 0 0/39 1/50 0/76 1/165 Hog Casing 25/350 0/24 2/20 0 27/394 Mouse Skin 0 0/43 0 0 0/43 Nitrile 1/359 0 0 0 1/359 Parafilm 0/112 1/36 3/71 0/94 4/313 Total 71/1,596 14/281 10/225 2/205 97/2,307

146

Table 6-4. Summary of the infection, dissemination, and transmission rates for the Culicoides insignis epizootic hemorrhagic disease virus (EHDV) vector competence trials. Two separate Florida populations were assayed. Trial # Population Provided Titer N IR (%)1 DR (%)2 TR (%)3 (PFU/mL) 1 Okeechobee 5.05 18 94.44 64.71 45.45 2 Okeechobee 3.03 64 21.88 7.14 100.00 3 Ocala 2.94 62 8.07 20.00 20.00 4 Ocala 4.00 93 4.30 50.00 0.00 1IR=Infection rate, calculated as a percent of the number of midges infected divided by the total sample size 2DR=Dissemination rate, calculated as a percent of the positive body samples divided by the number of infected midges 3TR=Transmission rate, calculated as a percent of the positive saliva samples divided by the number with disseminated infection

Figure 6-2. Viral titers in bodies, legs, and saliva for two populations of Culicoides insignis each at two inoculation titers. Error bars represent standard deviations for treatments that had greater than one positive individual.

147

LIST OF REFERENCES

Abubekerov, L.A., and B.A. Mullens. 2017. Egg and larval morphology of Culicoides sonorensis (Diptera: Ceratopogonidae). J. Med. Entomol. 55: 553-560. Adams, K.P., B.P. Murphy, and M.D. Ross. 2016. Captive white-tailed deer industry- Current status and growing threat. Wildlife Society Bulletin 40(1): 14-19. Agrela, I., E. Sanchez, B. Gomez, and M.D. Feliciangeli. 2002. Feeding behavior of Lutzomyia pseudolongipalpis (Diptera: Psychodidae), a putative vector of visceral leishmaniasis in Venezuela. Journal of Medical Entomology 39(3): 440-445. Akey, D.H., H.W. Potter, and R.H. Jones. 1978. Effects of rearing temperature and larval density on longevity, size, and fecundity in the biting gnat Culicoides variipennis. Ann. Entomol. Soc. Am. 71: 411-418. Allan, B.F., L.S. Goessling, G.A. Storch, and R.E. Thach. 2010. Blood meal analysis to identify reservoir hosts for Amblyomma americanum ticks. Emerg. Infect. Dis. 16(3): 433-440. Allison, A.B., V.H. Goekjian, A.C. Potgieter, W.C. Wilson, D.J. Johnson, P.P.C. Mertens, and D.E. Stallknecht. 2010. Detection of a novel reassortant epizootic hemorrhagic disease virus (EHDV) in the USA containing RNA segments derived from both exotic (EHDV-6) and endemic (EHDV-2) serotypes. J. Gen. Virol. 91: 430-439. Alto, B.W., K. Wiggins, B. Eastmond, D. Velez, L.P. Lounibos, and C.C. Lord. 2017. Transmission risk of two chikungunya lineages by invasive mosquito vectors from Florida and the Dominican Republic. PLoS Negl. Trop. Dis. 11:1-23. Alto, B.W., K. Wiggins, B. Eastmond, S. Ortiz, K. Zirbel, and L.P. Lounibos. 2018. Diurnal temperature range and chikungunya virus infection in invasive mosquito vectors. J. Med. Entomol. 55: 217-224. Ander, M., K. Troell, and J. Chirico. 2013. Barcoding of biting midges in the genus Culicoides: a tool for species determination. Med. Vet. Entomol. 27: 323-331. Anderson, D.P., B.J. Frosch, and J.L. Outlaw. 2007. Economic impact of the United States cervid farming industry. Agric. Food Policy Cent. Texas A&M Univ. Anderson, D.P., J.L. Outlaw, M.L. Earle, and J.W. Richardson. 2017. Economic impact of the U.S. deer breeding and hunting operations. Agric. Food Policy Cent. Texas A&M Univ: 1-20. Anderson, J.F., T.G. Andreadis, A.J. Main, and D.L. Kline. 2004. Prevalence of West Nile virus in tree canopy-inhabiting Culex pipiens and associated mosquitoes. Am. Soc. Trop. Med. Hyg. 71: 112-119.

148

Anderson, J.F., T.G. Andreadis, A.J. Main, F.J. Ferrandino, and C.R. Vossbrinck. 2006. West Nile virus from female and male mosquitoes (Diptera: Culicidae) in subterranean, ground, and canopy habitats in Connecticut. J. Med Entomol. 43: 1010-1019. Anderson, J.R. and A.X. Linhares. 1989. Comparison of several different trapping methods for Culicoides variipennis (Diptera: Ceratopogonidae). J. Am. Mosq. Control. Assoc. 5: 325-334. Anderson, S.L., S.L. Richards, and C.T. Smartt. 2010. A simple method for determining arbovirus transmission in mosquitoes. J. Am. Mosq. Control Assoc. 26: 108-111. Anishchenko M., R.A. Bowen, S. Paessler, L. Austgen, I.P. Greene, and S.C. Weaver. 2006. Venequelan encephalitis emergence mediated by a phylogenetically predicted viral mutation. Proc. Natl. Acad. Sci. U.S.A. 103: 4994-4999. Anthony, S.J., N. Maan, S. Maan, G. Sutton, H. Attoui, and P.P.C. Mertens. 2009a. Genetic and phylogenetic analysis of the core proteins VP1, VP3, VP4, VP6, and VP7 of epizootic haemorrhagic disease virus (EHDV). Virus Res. 145: 187-199. Anthony, S.J., S. Maan, N. Maan, L. Kgosana, K. Bachanek-Bankowska, C. Batten, K.E. Darpel, G. Sutton, H. Attoui, and P.P.C. Mertens. 2009b. Genetic and phylogenetic analysis of the outer-coat proteins VP2 and VP5 of epizootic haemorrhagic disease virus (EHDV): Comparison of genetic and serological data to characterize the EHDV serogroup. Virus Res. 145: 200-210. Aussel, J.P., and J.R. Linley. 1994. Natural food and feeding behavior of Culicoides furens larvae (Diptera, Ceratopogonidae). J. Med. Entomol. 31: 99-104. Bakhoum, M.T., M. Sarr, A.G. Fall, K. Huber, M. Fall, M. Sembène, M.T. Seck, K. Labuschagne, L. Gardès, M. Ciss, G. Gimonneau, J. Bouyer, T. Baldet, and C. Garros. 2018. DNA barcoding and molecular identification of field-collected Culicoides larvae in the Niayes area of Senegal. Parasit. Vectors. 11: 615. Barceló C. and M.A. Miranda. 2018. Bionomics of livestock-associated Culicoides (biting midge) bluetongue virus vectors under laboratory conditions. Med. Vet. Entomol. 32: 216-225. Barnard, B.J.H., G.H. Gerdes, and R. Meiswinkel. 1998. Some epidemiological and economic aspects of a bluetongue-like disease in cattle in South Africa-1995/96 and 1997. Onderstepoort J. Vet. Res. 65(3): 145-151. Battle, F. and E. Turner. 1972. Some nutritional and chemical properties of the larval habitats of certain species of Culicoides (Diptera: Ceratopogonidae). J. Med. Entomol. 9: 32-35.

149

Baylis, M., L. O’Connell, and P.S. Mellor. 2008. Rates of bluetongue virus transmission between Culicoides sonorensis and sheep. Med. Vet. Entomol. 22: 228-237. Becker, M.E., W.K. Reeves, S.K. Dejean, M.P. Emery, E.N. Ostlund, and L.D. Foil. 2010. Detection of bluetongue virus RNA in field-collected Culicoides spp. (Diptera: Ceratopogonidae) following the discovery of bluetongue virus serotype 1 in white- tailed deer and cattle in Louisiana. J. Med. Entomol. 47(2): 269-273. Becker, P. 1960. Observations on the life cycle and immature stages of Culicoides circumscriptus Kieff. (Diptera: Ceratopogonidae). Proc. R. Soc. Edinburgh. 67: 363- 386. Ben-Chetrit, E. and E. Schwartz. 2015. Vector-borne diseases in Haiti: A review. Travel Med. Infect. Dis. 13: 150-158. Bennett, G.F. 1960. On some ornithophilic blood-sucking Diptera in Algonquin Park, Ontario, Canada. Can. J. Zool. 38: 377-389. Bennett, K.E., K.E.N.E. Olson, M.D.L. Muñoz, I. Fernandez-Salas, J.A. Farfan-Ale, S. Higgs, W.C. Black IV, and B.J. Beaty. 2002. Variation in vector competence for dengue 2 virus among 24 collections of Aedes aegypti from Mexico and the United States. Am. J. Trop. Med. Hyg. 67(1): 85-92. Biggerstaff, B.J. 2003. Confidence intervals for the difference of two proportions estimated from pooled samples. J. Agric. Biol. Environ. Stat. 13: 478-496. Biggerstaff, B.J. 2009. PooledInfRate, Version 4.0: a Microsoft Office Excel Add-in to computer prevalence estimates from pooled samples. Blanton, F.S. and W.W. Wirth. 1979. The sand flies (Culicoides) of Florida (Diptera: Ceratopogonidae). Arthropods Florida Neighboring L. Areas 10:1-204. Blosser, E.M., T. Stenn, C. Acevedo, and N.D. Burkett-Cadena. 2016. Host use and seasonality of Culex (Melanoconion) iolambdis (Diptera: Culicidae) from eastern Florida, USA. Acta Trop. 164: 352-359. Boorman, J. 1987. Induction of salivation in biting midges and mosquitoes and demonstration of virus in the saliva of infected insects. Med. Vet. Entomol. 1: 211- 214. Borisevich V., A. Seregin, R. Nistler, D. Mutabazi, and V. Yamshchikov. 2006. Biological properties of chimeric West Nile viruses. Virology 349: 371-381. Borkent A. and W.L. Grogan. 2009. Catalog of the new world biting midges north of Mexico (Diptera: Ceratopogonidae). Zootaxa 48: 1-48. Borkent A. 1997. World species of biting midges (Diptera: Ceratopogonidae). Bull. of the AMNH 233.

150

Bowne, J.G. and R.H. Jones. 1966. Observations on bluetongue virus in the salivary glands of an insect vector, Culicoides variipennis. Virology 30: 127-133. Brault, A.C., A.M. Powers, E.C. Holmes, C.H. Woelk, and S.C. Weaver. 2002. Positively charged amino acid substitutions in the E2 envelope glycoprotein are associated with the emergence of Venequelan equine encephalitis virus. J. Virol. 76: 1718-1730. Brault, A.C., A.M. Powers, D. Ortiz, J.G. Estrada-franco, S.C. Weaver, B.J. Beaty, A.C. Brault, A.M. Powers, D. Ortiz, J.G. Estrada-franco, R. Navarro-lopez, and S.C. Weaver. 2018. Venezuelan equine encephalitis emergence: enhanced vector infection from a single amino acid substitution in the envelope glycoprotein. Proc. Nat. Acad. Sci. USA 101(31); 11344-11349. Braverman, Y., P.F.L. Boreham, and R. Galum. 1971. The origin of blood meals of female Culicoides pallidipennis trapped in a sheepfold in Israel. J. Med. Entomol. 8(4): 379-381. Braverman, Y. and P.E. Hulley. 1979. The relationship between the numbers and distribution of some antennal and palpal sense organs and host preference in some Culicoides (Diptera: Ceratogoponidae) from southern Africa. J. Med. Entomol. 15: 419-424. Brugger K., J. Köfer, and F. Rubel. 2016. Outdoor and indoor monitoring of livestock- associated Culicoides spp. to assess vector-free periods and disease risks. BMC Vet. Res. 12: 1-10. Brustolin, M., S. Talavera, A. Nuñez, C. santamaría, R. Rivas, N. Pujol, M. Valle, M. Verdún, A. Brun, N. Pagés, and N. Busquets. 2017. Rift Valley fever virus and European mosquitoes: Vector competence of Culex pipiens and Stegomyia albopicta (=Aedes albopictus). Med. Vet. Entomol. 31: 365-372. Burkett-Cadena, N.D., S.P. Graham, H.K. Hassan, C. Guyer, M.D. Eubanks, C.R. Katholi, and T.R. Unnasch. 2008. Blood feeding patterns of potential arbovirus vectors of the genus Culex targeting ectothermic hosts. Am. J. Trop. Med. Hyg. 79(5): 805-815. Bustin S.A. 2000. Absolute quantification of mRNA using real-time reverse transcription polymerase chain reaction assays. J. Mol. Endocrinol. 25: 169-193. Campbell, C.H., T.L. Barber, and M.M. Jochim. 1978. Antigenic relationship of Ibaraki, bluetongue, and epizootic hemorrhagic disease viruses. Vet. Microbiol 3: 15-22. Campbell, M.M. and D.S. Kettle. 1975. Oogenesis in Culicoides brevitarsis Kieffer (Diptera: Ceratopogonidae) and the development of a plastron-like layer on the egg. Aust. J. Zool. 23: 151-157.

151

Carpenter, S., A. Wilson, and P.S. Mellor. 2009. Culicoides and the emergence of bluetongue virus in northern Europe. Trends Microbiol 17: 172-178. Černý, O., J. Votýpka, and M. Svobodová. 2011. Spatial feeding preferences of ornithophilic mosquitoes, blackflies, and biting midges. Med. Vet. Entomol. 25: 104- 108. Chalmers, G., H. Vance, and G.J. Mitchell. 1964. An outbreak of epizootic hemorrhagic disease in wild ungulates in Alberta. Wild. Dis. 42: 15-20. Chaskopoulou, A., I.A. Giantsis, S. Demir, and M.C. Bon. 2016. Species composition, activity patterns, and blood meal analysis of sand fly populations (Diptera: Psychodidae) in the metropolitan region of Thessaloniki, an endemic focus of canine leishmaniasis. Acta Tropica 158: 170-176. Chaves, L.F., L.C. Harrington, C.L. Keogh, A.M. Nguyen, and U.D. Kitron. 2010. Blood feeding patterns of mosquitoes: random or structured? Frontiers in Zool. 7:3. Cheney, I.W., M. Yamakawa, P. Roy, J.O. Mecham, and W.C. Wilson. 1996. Molecular characterization of the segment 2 gene of epizootic hemorrhagic disease virus serotype 2: Gene sequence and genetic diversity. Virology 224: 555-560. Chiang, C.L. and W.C. Reeves. 1962. Statistical estimation of virus infection rates in mosquito vector populations. Am. J. Hyg. 75: 377-391. Colton, L., B.J. Biggerstaff, A. Johnson, and R.S. Nasci. 2005. Quantification of West Nile virus in vector mosquito saliva. J. Am. Mosq. Control. Assoc. 21: 49-53. Cortez, A.M., V.P.M. Silva, P.V.S. Queiroz, H.T.A. Andrade, M.I.B. Loiola, and M.F.F.M. Ximenes. 2007. Vertical stratification and development aspects of phlebotomine sand flies (Diptera: Psychodidae) in an area of Atlantic forest tree species in a metropolitan region in northeastern Brazil. J. Vector. Ecol 32: 336-341. Darbro, J.M. and L.C. Harrington. 2007. Avian defensive behavior and blood-feeding success of the West Nile vector mosquito, Culex pipiens. Behav. Ecol. 18(4): 750- 757. Davidson, W.R. (Ed.), F. Man. 2006. White-tailed Deer. In Field Manual of Wildlife Diseases in the Southeastern United States. Southeastern Cooperative Wildlife Disease Study. Dinesh, D.S., P. Das, A. Picado, C. Davies, N. Speybroeck, M. Boelaert, and M. Coosemans. 2008. The efficacy of indoor CDC light traps for collecting the sandfly Phlebotomus argentipes, vector of Leishmania donovani. Med. Vet. Entomol. 22: 120-123. Dove, W.E., D. G. Hall, and J.B. Hull. 1932. The salt marsh sand fly problem. Ann. Entomol. soc. Am. 25: 505-527.

152

Downes, J.A. 1958. The feeding habits of biting flies and their significance in classification. Annu. Rev. Entomol. 3: 249-266. Dubay,S.A., J.C. deVos, T.H. Noon, and S. Boe. 2004. Epizootiology of hemorrhagic disease in mule deer in central Arizona. J. Wildl. Dis. 40: 119-124. Ducheyne, E., R. De Deken, S. Bécu, B. Codina, K. Nomikou, O. Mangana- Vougiaki, G. Georgiev, B.V. Purse, and G. Hendickx. 2007. Quantifying the wind dispersal of Culicoides species in Greece and Bulgaria. Geospat. Health. 1: 177- 189. Dyce, A.L. and M.D. Murray. 1966. Notes on the pupal behavior and collection of larvae and pupae of some Australian Culicoides (Diptera: Ceratopogonidae). Aust. J. Entomol. 5:37-39. Dyce, A.L. 1969. The recognition of nulliparous and parous Culicoides (Diptera: Ceratopogonidae) without dissection. Aust. J. Entomol. 8: 11-15. Edman, J.D. and W.L. Bidlingmayer. 1964. Flight capcity of blood-engorged mosquitoes. Mosq. News 29: 386-392. Edman, J.D., L.A. Webber, and A.A. Schmid. 1974. Effect of host defenses on the feeding pattern of Culex nigripalpus when offered a choice of blood sources. J. Parasit. 60(5): 874-883. Elbers, A.R.W., A. Backx, K. Mintiens, G. Gerbier, C. Staubach, G. Hendrickx, and A. van der Spek. 2008. Field observations during bluetongue serotype 8 epidemic in 2006. II. Morbidity and mortality rate, case fatality and clinical recovery in sheep and cattle in the Netherlands. Prev. Vet. Med. 87: 31-40. Erram D.E. and N. Burkett-Cadena. 2018. Laboratory studies on the oviposition stimuli of Culicoides stellifer (Diptera: Ceratopogonidae), a suspected vector of Orbiviruses in the United States. Parasites & Vectors 11: 300. Estep, L.K., C.J.W. McClure, N.D. Burkett-Cadena, H.K. Hassan, T.L. Hicks, T.R. Unnasch, and G.E. Hill. 2011. A multi-year study of mosquito feeding patterns on avian hosts in a southeastern focus of eastern equine encephalitis virus. Am. J. Trop. Med. Hyg. 84: 718-726. Fabian, M.M., H. Toma, T. Arakawa, and Y. Sato. 2004. Malaria parasite developmental analyses by the nested polymerase chain reaction method: an implication for the evaluation of mosquito infection rates in epidemiological studies. SE Asian J .Trop. Med. Pub. Health 35(4): 820-827. Ferri, G., M. Alu, B. Corradini, M. Licata, and G. Beduschi. 2009. Species identification through DNA “barcodes”. Gen. Test. Mol. Biomark. 13(3): 421.

153

Flacke, G.L., M.J. Yabsley, B.A. Hanson, and D.E. Stallknecht. 2004. Hemorrhagic disease in Kansas: Enzootic stability meets epizootic disease. J. Wildl. Dis. 40: 288- 293. Foster, N.M., R.H. Jones, and B.R. McCrory. 1963. Preliminary investigations on insect transmission of bluetongue virus in sheep. Am. J. Vet. Res. 24(103): 1195- 1200. Foster, N.M., R.H. Jones, and A.J. Luedke. 1968. Transmission of attenuated and virulent bluetongue virus with Culicoides variipennis infected orally via sheep. J. Vet. Res. 19: 275-279. Foster, N.M. and R.H. Jones. 1973. Bluetongue virus transmission with Culicoides variipennis via embryonating chicken eggs. J. Med. Entomol. 10(6): 529-532. Foster, N.M., R.D. Breckon, A.J. Luedke, and R.H. Jones. 1977. Transmission of two strains of epizootic hemorrhagic disease virus in deer by Culicoides variipennis. J. Wildl. Dis. 13: 9-16. Fox, K.A., B. Diamond, F. Sun, A. Clavijo, L. Sneed, D.N. Kitchen, and L.L. Wolfe. 2015. Testicular lesions and antler abnormalities in Colorado, USA mule deer (Odocoileus hemionus): a possible role for epizootic hemorrhagic disease virus. J. Wild. Dis. 51: 166-176. Fu, H., C.J. Leake, P.P.C. Mertens, and P.S. Mellor. 1999. The barriers to bluetongue virus infection, dissemination and transmission in the vector, Culicoides variipennis (Diptera: Ceratopogonidae). Arch. Virol. 144: 747-761. Garrett-Jones, C. 1964. The human blood index of malarial vectors in relationship to epidemiological assessment. Bull. World Heal. Organ. Hlth. Org. 30: 241-261. Garros, C., L. Gardés, X. Alléne, I. Rakotoarivony, E. Viennet, S. Rossi, and T. Balenghien. 2011. Adaptation of a species-specific multiplex PCR assay for the identification of blood meal source in Culicoides (Ceratopogonidae: Diptera): applications on Palaearctic biting midge species, vectors of Orbiviruses. Infect. Gen. and Evol. 11: 1103-1110. Garvin, M.C. and E.C. Greiner. 2003. Ecology of Culicoides (Diptera: Ceratopogonidae) in southcentral Florida and experimental Culicoides vectors of the avian hematozoan Haemoproteus danilewskyi Kruse. J. Wildl. Dis. 39: 170-178. Gassett, J.W., D.P. Wiesler, A.G. Baker, D.A. Osborn, K.V. Miller, R.L. Marchinton, and M. Novotny. 1996. Volatile compounds from interdigital gland of male white- tailed deer (Odocoileus virginianus). J. Chem. Ecol. 22(9): 1689-1696.

154

Gassett, J.W., D.P. Wiesler, A.G. Baker, D.A. Osborn, K.V. Miller, R.L. Marchinton, and M. Novotny. 1997. Volatile compounds from the forehead region of male white-tailed deer (Odocoileus virginianus). Journal of Chemical Ecology 23(3): 569- 578. Gibbs, E.P.J. and M.J.P. Lawman. 1977. Infection of British deer and farm-animals with epizootic hemorrhagic disease of deer virus. J. Comp. Pathol. 87: 335-343. Greiner, E.C., E.S. Eveligh, and W.M. Boone. 1978. Ornithophilic Culicoides spp. (Diptera: Ceratopogonidae) from New Brunswick, Canada, and implications of their involvement in haemoproteid transmission. J. Med. Entomol. 14(6): 701-704. Greiner, E.C., T.L. Barber, J.E. Pearson, W.L. Kramer, and E.P. Gibbs. 1985. Orbiviruses from Culicoides in Florida. Prog. Clin. Biol. Res. 178: 195-200. Greiner, E.C., W.I. Knausenberger, M. Messersmith, W.L. Kramer, and E.P.J. Gibbs. 1990. Culicoides spp. (Diptera: Ceratopogonidae) associated with cattle in St. Croix, Virgin Islands, and their relevance to bluetongue virus. J. Med. Entomol. 27: 1071-1073. Grimes, J.M., J.N. Burroughs, P. Gouet, J.M. Diprose, R. Malby, S. Ziéntara, P.P.C. Mertens, and D.I. Stuart. 1998. The atomic structure of the bluetongue virus core. Nature 395: 470-478. Grogan Jr., W.L., L.J. Hribar, C.S. Murphree, and J.E. Cilek. 2010. New records of biting and predaceous midges from Florida, including species new to the fauna of the United States (Diptera: Ceratopogonidae). Insecta mundi 0147: 1-59. Gu, W., R. Lampman, and R.J. Novak. 2003. Problems in estimating mosquito infection rates using minimum infection rate. J. Med. Entomol. 40: 595-596. Guedes, D.R.D., M.H.S. Paiva, M.M.A. Donato, P.P. Barbosa, L. Krokovsky, S.W.S. Rocha, K.L.A. Saraiva, M.M. Crespo, T.M.T. Rezende, G.L. Wallau, R. Barbosa, C.M.F. Oliveira, M.A.V. Melo-santos, L. Pena, M.T. Cordeiro, R.F.D.O. Franca, A.L.S. De Oliveira, C.A. Peixoto, W.S. Leal, and C.F.J. Ayres. 2017. Zika virus replication in the mosquito Culex quinquefasciatus in Brazil. Emerg. Microb. & Inf. 6: e69. Haigh, J.C., C. Mackintosh, and F. Griffin. 2002. Viral, parasitic, and prion diseases of farmed deer. Rev. Sci. Tech.-Off. Int. des. Epizoot. 21: 219-248. Hair, J.A. and E.C. Turner Jr. 1968. Preliminary host preference studies on Virginia Culicoides (Diptera: Ceratopogonidae). Mosq. News 28(1): 103-107. Hall-Mendelin, S., S.A. Ritchie, C.A. Johansen, P. Zboroski, G. Cortis, S. Dandridge, R.A. Hall, and A.F. van den Hurk. 2010. Exploiting mosquito sugar feeding to detect mosquito-borne pathogens. Proc. Natl. Acad. Sci. 107: 11255- 11259.

155

Haouas, N., B. Pesson, R. Boudabous, J.P. Dedet, H. Babba, and C. Ravel. 2007. Development of a molecular tool for the identification of leishmania reservoir hosts by blood meal analysis in the insect vectors. Am. J. Trop. Med. Hyg. 77: 1054-1059. Hardy, J.L., E.J. Houk, L.D. Kramer, and W.C. Reeves. 1983. Intrinsic factors affecting vector comeptence of mosquitoes for arboviruses. Annu. Rev. Entomol. 28: 229-262. Harrup, L.E., S. Laban, B.V. Purse, Y.K. Reddy, Y.N. Reddy, S.M. Byregowda, N. Kumar, K.M. Purushotham, S. Kowalli, M. Prasad, G. Prasad, A.A. Bettis, R. De Keyser, J. Logan, C. Garros, D. Gopurenko, G. Bellis, K. bauschagne, B. Mathieu, and S. Carpenter. 2016. DNA barcoding and surveillance sampling strategies for Culicoides biting midges (Diptera: Ceratopogonidae) in southern India. Parasites and Vectors 9: 1-20. Harsha, R. and A. Mazumdar. 2015. Laboratory rearing of immature Culicoides peregrinus Kieffer, a potential vector of bluetongue virus. Med. Vet. Entomol. 29: 434-438. Harsha, R., N. Paul, G. Bellis, and A. Mazumdar. 2017a. Ultrastructure of the eggs of four species of Culicoides Latreille (Diptera: Ceratopogonidae) with a key to the eggs of all studied species by SEM. Zool. Anz. 269: 155-165. Harsha, R., N. Paul, and A. Mazumdar. 2017b. Description of immature stages of Culicoides peregrinus (Diptera: Ceratopogonidae), a potent vector of bluetongue virus. Turkish J. Zool. 41: 812-820. Henry, L.G. and T.R. Adkins Jr. 1975. Vertical distribution of biting midges in coastal South Carolina. Ann. Entomol. Soc. 68: 321-324. Hess, A.D., R.O. Hayes, and C.H. Tempelis. 1968. The use of the foraging ratio technique in mosquito host preference studies. Mosq. News 28: 386-389. Hess, A.D., and R.O. Hayes. 1970. Relative potentials of domestic animals for zooprophylaxis against mosquito vectors of encephalitis. Am. J. Trop. Med. Hyg. 19(2): 327-334. Hoff, G.L. and D.O. Trainer. 1973. Experimental infection in North American elk with epizootic hemorrhagic disease virus. J. Wildl. Dis. 9: 129-132. Holmes, E.C. and S.S. Burch. 2000. The causes and consequences of genetic variation in dengue virus. Trends Microbiol. 8: 74-77. Honorio, N.A., K. Wiggins, D.C. Camara, B. Eastmond, and B.W. Alto. 2018. Chikungunya virus vector competency of Brazilian and Florida mosquito vectors. PLoS Negl. Trop. Dis. 12: 1-17. Huismans, H., C.W. Bremer, and T.L. Barber. 1979. The nucleic acid and proteins of epizootic haemorrhagic disease virus. Onderstepoort J. Vet. Res. 46: 95-104.

156

Huismans, H. and H.J. Els. 1979. Characterization of the tubules associated with the replication of three different orbiviruses. Virology 92: 397-406. Humphreys, J.G. and E.C. Turner, Jr. 1973. Blood-feeding activity of female Culicoides (Diptera: Ceratopogonidae). J. Med. Entomol. 10: 79-83. Hunt, G.J. and C.N. McKinnon. 1990. Evaluation of membranes for feeding Culicoides variipennis (Diptera: Ceratopogonidae) with an improved artificial blood-feeding apparatus. J. Med. Entomol. 27: 934-937. Hyatt, A.D., A.R. Gould, B. Coupar, and B.T. Eaton. 1991. Localization of the non- structural protein NS3 in bluetongue virus-infected cells. J. Gen. Virol. 72: 2263- 2267. Hyatt, A.D., Y. Zhao, and P. Roy. 1993. Release of bluetongue virus-like particles from insect cells is mediated by BTV nonstructural proteins NS3/NS3A. Virology 193: 592-603. Isberg, E., Y. Hillbur, and R. Ignell. 2013. Comparative study of antennal and maxillary palp olfactory sensilla of female biting midges (Diptera: Ceratopogonidae: Culicoides) in the context of host preference and phylogeny. J. Med. Entomol. 50: 485-492. Jamnback, H. 1965. The Culicoides of New York state (Diptera: Ceratopogonidae). New York State Mus. Sci. Serv. Bull. Albany, New York. Jennings, D.M. and P.S. Mellor. 1987. Variation in the responses of Culicoides variipennis (Diptera: Ceratopogonidae) to oral infection with bluetongue virus. Arch. Virol. 95: 177-182. Jones, R.H. 1961. Observations on the larval habitats of some North American species of Culicoides (Diptera: Ceratopogonidae). Ann. Entomol. Soc. Am. 54: 702-710. Jones, R.H., H.W. Potter, and S.K. Baker. 1969. An improved larval medium for colonized Culicoides variipennis. J. Econ. Entomol. 62: 1483-1486. Jones, R.H., R.D. Roughton, N.M. Foster, and B.M. Bando. 1977. Culicoides, the vector of epizootic hemorrhagic disease in white-tailed deer in Kentucky in 1971. J. Wildl. Dis. 13: 2-8. Jones, R.H. and N.M. Foster. 1978. Heterogeneity of Culicoides variipennis field populations to oral infection with bluetongue virus. Am. J. Trop. Med. Hyg. 27: 178- 183. Jones, R.H., E.T. Schmidtmann, and N.M. Foster. 1983. Vector-competence studies for bluetongue and epizootic hemorrhagic disease viruses with Culicoides venustus (Ceratopogonidae). Mosq. News. 43: 184-186.

157

Kaufmann, C., A. Mathis, and C. Vorburger. 2015. Sugar-feeding behavior and longevity of European Culicoides biting midges. Med. Vet. Entomol. 29: 17-25. Kedmi, M., Y. Herziger, N. Galon, R.M. Cohen, M. Perel, C. Batten, Y. Braverman, Y. Gottlieb, N. Schpigel, and E. Klement. 2010a. The association of winds with the spread of EHDV in dairy cattle in Israel during an outbreak in 2006. Prev. Vet. Med. 96: 152-160. Kedmi, M., M. Van Straten, E. Ezra, N. Galon, and E. Klement. 2010b. Assessment of the productivity effects associated with epizootic hemorrhagic disease in dairy herds. J. Dairy Sci. 93: 2486-2495. Kettle, D.S. and J.W. Lawson. 1952. The early stages of British biting midges Culicoides Latreille (Diptera: Ceratopogonidae) and allied genera. Bull. Entomol. Res. 43: 421-467. Kirstein, F. and J.S. Gray. 1996. A molecular marker for the identification of the zoonotic reservoirs of lyme borreliosis by analysis of the blood meal in its European vector Ixodes ricinus. Appl. Environ. Microbiol. 62: 4060-4065. Kluiters, G, H. Swales, and M. Baylis. 2015. Local dispersal of palearctic Culicoides biting midges estimated by mark-release-recapture. Parasites and Vectors 8: 86. Koch, H.G. and R.C. Axtell. 1978. Autogeny and rearing of Culicoides furens, C. hollensis, and C. melleus (Diptera: Ceratopogonidae) from coastal North Carolina. Mosq. News. 38: 240-244. Koch, H.G. and R.C. Axtell. 1979. Attraction of Culicoides furens and C. hollensis (Diptera: Ceratopogonidae) to animal hosts in a salt marsh habitat. J. Med. Entomol. 15: 494-499. Kramer, W.L., E.C. Greiner, and E.P.J. Gibbs. 1985a. A survey of Culicoides midges (Diptera: Ceratopogonidae) associated with cattle operations in Florida, USA. J. Med. Entomol. 22: 153-162. Kramer, W.L., E.C. Greiner, and E.P.J. Gibbs. 1985b. Seasonal variations in population size, fecundity, and parity rates of Culicoides insignis (Diptera: Ceratopogonidae) in Florida, USA. J. Med. Entomol. 22: 163-169. Kramer, L.D. and A.T. Ciota. 2015. Dissecting vectorial capacity for mosquito-borne viruses. Curr. Opin. Virol. 15: 112-118. Lardeux, F., P. Loayza, B. Bouchité, and T. Chavez. 2007. Host choice and human blood index of Anopheles pseudopunctipennis in a village of the Andean valleys of Bolivia. Malaria J. 6: 8. Lassen, S.B., S.A. Nielsen, and M. Kristensen. 2012. Identity and diversity of blood meal hosts of biting midges (Diptera: Ceratopogonidae: Culicoides Latreille) in Denmark. Parasites and Vectors 5: 143-151.

158

Lechowicz, M.J. 1982. The sampling characteristics of electivity indices. Oecologia 52(1): 22-30. Lee, J. and A. English. 2011. Economic impact of the deer & elk farming industry in Indiana. Indiana Deer Elk Farming Assoc. Li, C., Z. Jiang, Y. Zeng, and Z. You. 2007. A note on environmental elements as essential prerequisites for behavioral expresion: A case study of Pére David’s deer. App. Anim. Behav. Sci. 103: 174-180. Liames, J.S., E. Cooter, D. Schwede, and J. Williams. 2018. A comparison of simulated and field-derived leaf area index (LAI) and canopy height values from four forest complexes in the southeastern USA. Forests 9: 1-20. Lillie, T.H., W.C. Marquardt, and R.H. Jones. 1981. The flight range of Culicoides variipennis (Diptera: Ceratopogonidae). Can. Entomol. 113: 419-426. Lillie, T.H., D.L. Kline, and D.W. Hall. 1985. The dispersal of Culicoides mississippiensis (Diptera: Ceratopogonidae) in a salt marsh near Yankeetown, Florida. J. Am. Mosq. Control. Assoc. 1: 463-467. Lim S.M., P. Koraka, S. van Boheemen, J.M. Roose, D. Jaarsma, D.A.M.C. van de Vijver, A.D.M.E. Osterhaus, and B.E.E. Martina. 2013. Characterization fo the mouse neuroinvasiveness of selected European strains of West Nile virus. PLoS One 8(9): e745765. Linley, J.R. 1969. Studies on larval development in Culicoides furens (Poey) (Diptera: Ceratopogonidae). I. Establishment of a standard rearing technique. Ann. Entomol. Soc. Am. 62: 702-711. Linley, J.R. 1983. Autogeny in the Ceratopogonidae: Literature and notes. Florida Entomol. 66: 228-234. Linley, J.R. and D.S. Kettle. 1964. A description of the larvae and pupae of Culicoides furens Poey, and Culicoides hoffmani fox (Diptera: Ceratopogonidae). Ann. Mag. Nat. Hist. 7: 129-149. Lysyk, T.J. and T. Danyk. 2007. Effect of temperature on life history parameters of adult Culicoides sonorensis (Diptera: Ceratopogonidae) in relation to geographic origin and vectorial capacity for bluetongue virus. J. Med. Entomol. 44: 741-751. Maan, N.S., S. Maan, A.C. Potgieter, I.M. Wright, M. Belaganahalli, and P.P.C. Mertens. 2017. Development of real-time RT-PCR assays for detection and typing of epizootic haemorrhagic disease virus. Transbound. Emerg. Dis. 64: 1120-1132. MacDonald, G. 1957. The epidemiology and control of malaria. Oxford University Press: London.

159

Magnarelli, L.A. 1981. Parity, follicular development, and sugar feeding in Culicoides melleus and C. hollensis. Environ. Entomol. 10: 807-811. Maguire, D.Y., K. Robert, K. Brochu, M. Larrivée, C.M. Buddle, and T.A. Wheeler. 2014. Vertical stratification of beetles (Coleoptera) and flies (Diptera) in temperate forest canopies. Environ. Entomol. 43:9-17. Mahmood, F., R.E. Chiles, Y. fang, and W.K. Reisen. 2004. Methods for studying the vector comeptence of Culex tarsalis for western equine encephalomyelitis virus. J. Am. mosq. Control. Assoc. 20: 277-282. Manly, B.F., L. McDonald, and D. Thomas. 2007. Resource selection by animals: statistical design and analysis for field studies, 2nd Ed. Dordrecht, The Netherlands: Kluwer Academic Publishers, 50-62. Mazumdar, S.M. and A. Mazumdar. 2016. Culicoides spp. Latreille (Diptera: Ceratopogonidae) vector of several human diseases-a neglected issue. NBU J. Anim. Sci. 10: 21-30. McDermott, E.G. and B.A. Mullens. 2014. Dessication tolerance in the eggs of the primary North American bluetongue virus vector Culicoides sonorensis (Diptera: Ceratopogonidae) and implications for vector persistence. J. Med. Entomol. 51: 1151-1158. McDermott, E.G., C.E. Mayo, A.C. Gerry, D. Laudier, N.J. MacLachlan, and B.A. Mullens. 2015. Bluetongue virus infection creates light averse Culicoides vectors and serious errors in transmission risk estimates. Parasites and Vectors 8: 1-9. McDermott, E.G., C.E. Mayo, B.A. Mullens, and T. Lysyk. 2017. Low temperature tolerance of Culicoides sonorensis (Diptera: Ceratopogonidae) eggs, larvae, and pupae from temperate and subtropical climates. J. Med. Entomol. 54: 264-274. McGregor, B.L., A.E. Runkel IV, S.M. Wisely, and N.D. Burkett-Cadena. 2018a. Vertical stratificiation of Culicoides biting midges at a Florida big game preserve. Parasites and Vectors 11: 505. McGregor, B.L., T. Stenn, A. Sayler, E.M. Blosser, J.K. Blackburn, S.M. Wisely, and N.D. Burkett-Cadena. 2018b. Host use patterns of Culicoides spp. biting midges at a big game preserve in Florida, U.S.A., and implications for the transmission of orbiviruses. Med. Vet. Entomol. In Print. McIntyre, M.P. 1976. The potential for commercial deer farming in New Zealand. Calif. Geogr. 16: 19-29. Mecham, J.O. and V.C. Dean. 1988. Protein coding assignment for the genome of epizootic haemorrhagic disease virus. J. Gen. Virol. 69: 1255-1262.

160 de Meillon, B. and W.W. Wirth. 1991. The genera and subgenera (excluding Culicoides) of the Afrotropical biting midges (Diptera: Ceratopogonidae). Ann. Natal Museum. 32: 27-147. Mellor, P.S. 1990. The replication of bluetongue virus in Culicoides vectors, pp. 143- 161. In Curr. Top. Microbiol. Immunol. Mellor, P.S., J. boorman, and M. Baylis. 2000. Culicoides biting midges: Their role as arbovirus vectors. Annu. Rev. Entomol. 45: 307-340. Mellor, P.S. and C. Hamblin. 2004. African horse sickness. Vet. Res. 35: 445-466. Mertens, P.P.C., S.P. Cowley, J.N. Burroughs, A.H. Corteyn, M.H. Jeggo, D.M. Jennings, and B.M. Gorman. 1989. Analysis of the roles of bluetongue virus outer capsid proteins VP2 and VP5 in determination of virus serotype. Virology 170: 561- 565. Mills, M.K., M.G. Ruder, D. Nayduch, K. Michel, and B.S. Drolet. 2017. Dynamics of epizootic hemorrhagic disease virus infection within the vector, Culicoides sonorensis (Diptera: Ceratopogonidae). PLoS One 12: 1-16. Molaei, G. and T.G. Andreadis. 2006. Identification of avian and mammalian derived bloodmeals in Aedes vexans and Culiseta melanura (Diptera: Culicidae) and its implication for West Nile virus transmission in Connecticut, U.S.A. J. Med. entomol 432: 1088-1093. Molaei, G. A. Farajollahi, P.M. Armstrong, J. Oliver, J.J. Howard, and T.G. Andreadis. 2009. Identification of bloodmeals in Anopheles quadrimaculatus and Anopheles punctipennis from eastern equine encephalitis virus foci in northeastern U.S.A. Med. Vet. Entomol 23: 350-356. Moore, D.L. and V.H. Lee. 1972. Antigenic relationship between the virus of epizootic haemorrhagic disease of deer and bluetongue virus. Archiv. für die gesamte Virusforschung 37: 282-284. Mooring, M.S., D.T. Blumstein, D.D. Reisig, E.R. Osborne, and J.M. Niemeyer. 2007. Insect-repelling behavior in bovids: role of mass, tail length, and group size. Biol. J. Linn. Soc. 91: 383-392. Mullen, G., M.E. Hayes, and K.E. Nusbaum. 1985a. Potential vectors of bluetongue and epizootic hemorrhagic disease viruses of cattle and white-tailed deer in Alabama. Prog. Clin. Biol. Res. 178: 201-206. Mullen, G.R., R.H. Jones, Y. Braverman, and K.E. Nusbaum. 1985b. Laboratory infections of Culicoides debilipalpis and C. stellifer (Diptera: Ceratopogonidae) with bluetongue virus. Prog. Clin. Biol. Res. 178: 239-243.

161

Mullens, B.A. and E.T. Schmidtmann. 1982. The gonotrophic cycle of Culicoides variipennis (Diptera: Ceratopogonidae) and its implications in age-grading field populations in New York state, USA. J. Med. Entomol. 19: 340-349. Mullens, B. and D. Rutz. 1983. Development of immature Culicoides variipennis (Diptera: Ceratopogonidae) at constant laboratory temperatures. J. Med. entomol. 76: 747-751. Mullens, B.A. 1985. Age-related adult activity and sugar feeding by Culicoides variipennis (Diptera: Ceratopogonidae) in southern California, USA. J. Med. Entomol. 22: 32-37. Mullens, B.A. and R.K. Velten. 1994. Rearing Culicoides variipennis sonorensis (Diptera, Ceratopogonidae) on agar and nematodes. J. Med. entomol. 31: 175-177. Muller, M.J. 1987. Transmission and in vitro extretion of bluetongue virus serotype 1 by inoculated Culicoides brevitarsis (Diptera: Ceratopogonidae). J. Med. Entomol. 24: 206-211. Müller-Schwarze, D., R. Altieri, and N. Porter. 1984. Alert odor from skin gland in deer. J. Chem. Ecol. 10(12): 1707-1729. Munga, S., N. Minakawa, G. Zhou, E. Mushinzimana, O.O.J. Barrack, A.K. Githeko, and G. Yan. 2006. Association between land cover and habitat productivity of malaria vectors in western Kenyan highlands. Am. J. Trop. Med. Hyg. 74(1): 69-75. Murphree, C.S. and G.R. Mullen. 1991. Comparative larval morphology of the genus Culicoides Latreille (Diptera: Ceratopogonidae) in North America with a key to species. Bull. Soc. Vector. Ecol. 16: 269-399. Murphy, M.D., B.A. Hanson, E.W. Howerth, and D.E. Stallknecht. 2006. Molecular characterization of epizootic hemorrhagic disease virus serotype 1 associated with a 1999 epizootic in white-tailed deer in the eastern United States. J. Wildl. Dis. 42. 616-624. Muturi, C.N., J.O. Ouma, I.I. Malele, R.M. Ngure, J.J. Rutto, K.M. Mithöfer, J. Enyaru, and D.K. Masiga. 2011. Tracking the feeding patterns of tsetse flies (Glossina genus) by analyis of bloodmeals using mitochondrial cytochrome genes. PLoS One: doi: 0017284. Nel, L. and H. Huismans. 1991. Synthesis of the virus-specified tubules of epizootic haemorrhagic disease virus using a baculovirus expression system. Virus. Res. 19: 139-152.

162

Nol, P., C. Kato, W.K. Reeves, J. Rhyan, T. Spraker, T. Gidlewski, K. VerCauteren, and M. Salman. 2010. Epizootic hemorrhagic disease outbreak in a captive facility housing white-tailed deer (Odocoileus virginianus), bison (Bison bison), elk (Cervus elaphus), cattle (Bos taurus), and goats (Capra hircus) in Colorado, USA. J Zoo Wildl Med. 41(3): 510-515. Nusinovici, S., H. Seegers, A. Joly, F. Beaudeau, and C. Fourichon. 2012. Increase in the occurrence of abortions associated with exposure to the bluetongue virus serotype 8 in naïve dairy herds. Theriogenology 78: 1140-1151. Ohashi, S., K. Yoshida, Y. Watanabe, and T. Tsuda. 1999. Identification and PCR- restriction fragment length polymorphism analysis of a variant of the Ibaraki virus from naturally infected cattle and aborted fetuses in Japan. J. Clin. Microbiol 37: 3800-3803. Ohashi, S., K. Yoshida, T. Yanase, and T. Tsuda. 2002. Analysis of intratypic variation evident in an Ibaraki virus strain and its epizootic hemorrhagic disease virus serogroup. J. Clin. Microbiol. 40: 3684-3688. Omori, T., Y. Inaba, T. Morimoto, Y. Tanaka, R. Ishitani, H. Kurogi, K. Munakata, K. Matsuda, and M. Matumoto. 1970. Ibaraki virus, an agent of epizootic disease of cattle resembling bluetongue: IV. Physicochemical and serological properties of the virus. Jpn. J. Microbiol. 14: 351-360. Pagés, N., S. Talavera, M. Verdún, N. Pujol, M. Valle, A. Bensaid, and J. Pujols. 2018. Schmallenberg virus detection in Culicoides biting midges in Spain: First laboratory evidence for highly efficient infection of Culicoides of the Obsoletus complex and Culicoides imicola. Transbound. Emerg. Dis. 65: e1-e6. Parida, S.K., R.K. Hazra, N. Marai, H.K. Tripathy, and N. Mahapatra. 2006. Host feeding patterns of malaria vectors of Orissa, India. J. Am. Mosq. Control. Assoc. 22(4): 629-634. Parker, G.G. 1995. Structure and microclimate of forest canopies. In M.D. Lowman, & N.M. Nadkarni (Eds.), Forest Canopies, pp 73-106. San Diego: Academic Press. Parsonson, I. and W. Snowdon. 1985. Bluetongue, epizootic hemorrhagic disease of deer and related viruses: current situation in Australia. Prog. Clin. Biol. Res. 178: 27-35. Paweska, J.T., G.J. Venter, P. Guillet, and C. Hamblin. 2005. A comparison of the susceptibility of Culicoides imicola and C. bolitinos to oral infection with eight serotypes of epizootic haemorrhagic disease virus. Med. Vet. Entomol. 19: 200- 207. Pearse, A.J. 1990. What has made deer farming in New Zealand so successful? Rangifer 5: 6-13.

163

Pfannenstiel, R.S., B.A. Mullens, M.G. Ruder, L. Zurek, L.W. Cohnstaedt, and D. Nayduch. 2015. Management of North American Culicoides biting midges: Current knowledge and research needs. Vector-borne Zoon. Dis. 15: 374-384. Ponlawat, A. and L.C. Harrington. 2005. Blood feeding patterns of Aedes aegypti and Aedes albopictus in Thailand. J. Med. Entomol. 42(5): 844-849. Prestwood, A.K., T.P. Kistner, F.E. Kellogg, and F.A. Hayes. 1974. The 1971 outbreak of hemorrhagic disease among white-tailed deer of the southeastern United States. J. Wildl. Dis. 10: 217-224. Purse, B.V., S. Carpenter, G.J. Venter, G. Bellis, and B.A. Mullens. 2015. Bionomics of temperate and tropical Culicoides midges: Knowledge gaps and consequences for transmission of Culicoides-borne viruses. Annu. Rev. Entomol. 60: 373-392. Qi, W.H., J. Li, X.Y. Zhang, Z.K. Wang, X.X. Li, C.Z. Yang, W.I. Fu, and B.S. Yue. 2011. The reproductive performance of female forest musk deer (Moschus berezovskii) in captivity. Theriogenology 76: 874-881. R Core Team. 2016. R: A language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria. Raich, T., M. Jacobson, F. Holbrook, R. Babion, C. Blair, and B. Beaty. 1997. Culicoides variipennis (Diptera: Ceratogoponidae) host selection in Colorado. J. Med. Entomol. 34: 247-249. Reeves, W.K. 2008. Osmoregulatory organs of immature Culicoides sonorensis (Diptera: Ceratopogonidae) in North America. Entomol. News. 119: 371-374. Richards S.L., S.L. Anderson, and C.C. Lord. 2014. Vector competence of Culex pipiens quinquefasciatus (Diptera: Culicidae) for West Nile virus isolates from Florida. Trop. Med. Int. Health 19: 610-617. Rigot, T., M. Vercauteren Drubbel, J.C. Delécolle, and M. Gilbert. 2013. Farms, pastures, and woodlands: The fine-scale distribution of palearctic Culicoides spp. biting midges along an agro-ecological gradient. Med. Vet. Entomol. 27: 29-38. Romero-Alvarez, D. and L.E. Escobar. 2017. Oropouche fever, an emergent disease from the Americas. Microbes Infect. 1-12. Rossi, E., G. Bongiorno, E. Ciolli, T. Di Muccio, A. Scalone, M. Gramiccia, L. Gradoni, and M. Maroli. 2007. Seasonal phenology, host-blood feeding preferences and natural Leishmania infection of Phlebotomus perniciosus (Diptera, Psychodidae) in a high-endemic focus of canine leishmaniasis in Rome province, Italy. Acta Tropica 105: 158-165.

164

Ruder, M.G., E.W. Howerth, D.E. Stallknecht, A.B. Allison, D.L. Carter, B.S. Drolet, E. Klement, and D.G. Mead. 2012. Vector competence of Culicoides sonorensis (Diptera: Ceratopogonidae) to epizootic hemorrhagic disease virus serotype 7. Parasites and Vectors 5: 236. Ruder, M.G., T.J. Lysyk, D.E. Stallknecht, L.D. Foil, D.J. Johnson, C.C. Chase, D.A. Dargatz, and E.P.J. Gibbs. 2015a. Transmission and epidemiology of bluetongue and epizootic hemorrhagic disease in North America. Current perspectives, research gaps, and future directions. Vector-borne Zoo. Dis. 15: 348-363. Ruder, M.G., D.E. Stallknecht, E.W. Howerth, D.L. Carter, R.S. Pfannenstiel, A.B. Allison, and D.G. Mead. 2015b. Effect of temperature on replication of epizootic hemorrhagic disease viruses in Culicoides sonorensis (Diptera: Ceratopogonidae). J. Med. Entomol. 52: 1050-1059. Ruder, M.G., D.E. Stallknecht, A.B. Allison, D.G. Mead, D.L. Carter, and E.W. Howerth. 2016. Host and potential vector susceptibility to an emerging orbivirus in the United States: Epizootic hemorrhagic disease virus serotype 6. Vet. Path. 53(3): 574-584. Rushton, J. and N. Lyons. 2015. Economic impact of bluetongue: a review of the effects on production. Veterinaria Italiana 51(4): 401-406. Saegerman, C., D. Berkvens, and P.S. Mellor. 2008. Bluetongue epidemiology in the European Union. Emerg. Infect. Dis. 14: 539-544. Saenz, M.R. and E.C. Greiner. 1994. Culicoides aspirated from cattle in Costa Rica, Honduras, Panama and Puerto Rico, and their role as potential vectors of bluetongue viruses. Med. Vet. Entomol. 8: 15-19. Samuel, P.P., N. Arunachalam, J. Hiriyan, and B.K. Tyagi. 2008. Host feeding patterns of Japanese encephalitis virus vector mosquitoes (Diptera: Culicidae) from Kuttanadu, Kerala, India. J. Med. Entomol. 45(5): 927-932. Sanders, C.J., R. Selby, S. Carpenter, and D.R. Reynolds. 2011. High-altitude flight of Culicoides biting midges. Vet. Rec. 169: 208. Santiago-Alarcon, D., P. Havelka, H.M. Schaefer, and G. Segelbacher. 2012. Bloodmeal analysis reveals avian Plasmodium infections and broad host preferences of Culicoides (Diptera: Ceratopogonidae) vectors. PLoS ONE 7(2): e31098. Santiago-Alarcon, D., P. Havelka, E. Pineda, G. Segelbacher, and H.M. Schaefer. 2013. Urban forests as hubs for novel zoonosis: blood meal analysis, seasonal variation in Culicoides (Diptera: Ceratopogonidae) vectors, and avian haemosporidians. Parasitol. 140: 1799-1810.

165

Santman-Berends, I.M.G.A., J.J. Hage, T.J.G.M. Lam, O.C. sampimon, and G. van Schaik. 2011. The effect of bluetongue virus serotype 8 on milk production and somatic cell count in Dutch dairy cows in 2008. J. Dairy Sci. 94: 1347-1354. Sarri, C., C. Stamatis, T. Sarafidou, I. Galara, V. Godosopoulos, M. Kolovos, C. Liakou, S. Tastsoglou, and Z. Mamuris. 2014. A new set of 16s rRNA universal primers for identification of animal species. Food Cont. 43: 35-41. Savini, G., A. Afonso, P. Mellor, I. Aradaib, H. Yadin, M. Sanaa, W. Wilson, F. Monaco, and M. Domingo. 2011. Epizootic hemorrhagic disease. Res. Vet. Sci. 91: 1-17. Schmidtmann, E.T., C.J. Jones, and B. Gollands. 1980. Comparative host-seeking activity of Culicoides (Diptera: Ceratopogonidae) attracted to pastured livestock in central New York state, USA. J. Med. Entomol. 17: 221-231. Schmidtmann, E.T., R.J. Bobian, and R. Belden. 1996. Soil chemistries define aquatic habitats with immature populations of Culicoides variipennis complex (Diptera: Ceratopogonidae). J. Med. Entomol. 37: 58-64. Schmidtmann, E.T., M.V. Herrero, A.L. Green, D.A. Dargatz, J.M. Rodriguez, and T.E. Walton. 2011. Distribution of Culicoides sonorensis (Diptera: Ceratopogonidae) in Nebraska, South Dakota, and North Dakota: Clarifying the epidemiology of bluetongue disease in the northern great plains region of the United States. J. Med. Entomol. 48: 634-643. Service, M.W. 1971. Adult flight activities of some British Culicoides species. J. Med Entomol. 8: 605-609. Sherich, K., A. Pocewicz, and P. Morgan. 2007. Canopy characteristics and growth rates of ponderosa pine and Douglas-fir at long established forest edges. Can. J. For. Res. 37: 2096-2105. Shope, R., L MacNamara, and R. Mangold. 1955. Report on the deer mortality, epizootic hemorrhagic disease of deer. N. J. Outdoors 6: 17-21. Slama, D., N. Haouas, H. Mezhoud, H. Babba, and E. Chaker. 2015. Blood meal analysis of Culicoides (Diptera: Ceratopogonidae) in Central Tunisia. PLoS One: doi: 0120528. Sleeman, J.M., J.E. Howell, W. Matthew Knox, and P.J. Stenger. 2009. Incidence of hemorrhagic disease in white-tailed deer is associated with winter and summer climatic conditions. ecohealth 6: 11-15. Sloyer, K.E., S.M. Wisely, and N.D. Burkett-Cadena. 2018. Effects of ultraviolet LED versus incandescent bulb and carbon dioxide for sampling abundance and diversity of Culicoides in Florida. J. Med. Entomol. In Print.

166

Smith, K.E. and D.E. Stallknecht. 1996. Culicoides (Diptera: Ceratopogonidae) collected during epizootics of hemorrhagic disease among captive white-tailed deer. J. Med. Entomol. 33: 507-510. Smith, K.E., D.E. Stallknecht, and V.F. Nettles. 1996a. Experimental infection of Culicoides lahillei (Diptera: Ceratopogonidae) with epizootic hemorrhagic disease virus serotype 2 (Orbivirus: Reoviridae). J. Med. Entomol. 33: 117-122. Smith, K., D. Stallknecht, C. Sewell, E. Rollor, G. Mullen, and R. Anderson. 1996b. Monitoring of Culicoides spp. at a site enzootic for hemorrhagic disease in white- tailed deer in Georgia, USA. J. Wildl. Dis. 32: 627-642. Snow, W.E. 1955. Feeding activities of some blood-sucking Diptera with reference to vertical distribution in bottomland forest. Ann. Entomol. Soc. Am. 48: 512-521. Stallknecht, D.E., J.L. Blue, E.A. Rollor III, V.F. Nettles, W.R. Davidson, and J. Pearson. 1991. Precipitating antibodies to epizootic hemorrhagic disease and bluetongue viruses in white-tailed deer in the southeastern United States. J. Wildl. Dis. 27: 238-247. Stallknecht, D.E., M.P. Luttrell, K.E. Smith, and V.F. Nettles. 1996. Hemorrhagic disease in white-tailed deer in Texas: A case for enzootic stability. J. Wildl. Dis. 32: 695-700. Stallknecht, D. and E. Howerth. 2004. Epidemiology of bluetongue and epizootic haemorrhagic disease in wildlife: Surveillance methods. Vet. Ital. 40: 203-207. Stallknecht, D.E., A.B. Allison, A.W. Park, J.E. Phillips, V.H. Goekjian, V.F. Nettles, and J.R. Fischer. 2015. Apparent increase of reported hemorrhagic disease in the Midwestern and Northeastern USA. J. Wildl. Dis. 51: 348-361. Steiger, D.M., P. Johnson, D.W. Hilbert, S. Ritchie, D. Jones, and S.G. Laurance. 2012. Effects of landscape disturbance on mosquito community composition in tropical Australia. J. Vect. Ecol. 37(1): 69-76. Stevens, G., B. McCluskey, A. King, E. O’Hearn, and G. Mayr. 2015. Review of the 2012 epizootic hemorrhagic disease outbreak in domestic ruminants in the United States. PLoS One 10: 1-12. Stewart, R.G. and D.L. Kline. 1999. Sugar feeding by Culicoides mississippiensis (Diptera: Ceratopogonidae) on the yaupon holly, Ilex vomitoria. J. Med. Entomol. 36: 268-271. Sun, W.K.C. 1974. Laboratory colonization of biting midges (Diptera: Ceratopogonidae). J. Med. Entomol. 11: 71-73. Swanson, D.A. and P.H. Adler. 2010. Vertical distribution of haematophagous Diptera in temperate forests of the southeastern U.S.A. Med. Vet. Entomol. 24: 182-188.

167

Swanson, D.A., P.H. Adler, and B. Malmqvist. 2012. Spatial stratification of host- seeking Diptera in boreal forests of northern Europe. Med. Vet. Entomol. 26: 56-62. Synek, P., P. Munclinger, T. Albrecht, and J. Votypka. 2013. Avian haemosporidians in haematophagous insects in the Czech Republic. Parasitol. Res. 112: 839-845. Szmaragd, C., A. Wilson, S. Carpenter, P.P.C. Mertens, P.S. Mellor, and S. Gubbins. 2007. Mortality and case fatality during the recurrence of BTV-8 in northern Europe in 2007. Vet. Rec. 571-572. Talavera, S., F. Muñoz-Muñoz, M. Verdún, and N. Pagés. 2017. Morphology and DNA barcoding reveal three species in one: description of Culicoides cryptipulicaris sp. nov. and Culicoides quasipulicaris sp. nov. in the subgenus Culicoides. Med. Vet. Entomol. 31: 178-191. Tanner, G.D. and J.E.C. Turner. 1974. Vertical activities and host preferences of several Culicoides species in a southwestern Virginia forest. Mosq. News. 34: 66- 70. Tanya, V.N., E.C. Greiner, and E.P.J. Gibbs. 1992. Evaluation of Culicoides insignis (Diptera: Ceratopogonidae) as a vector of bluetongue virus. Vet. Microbiol. 32: 1- 14. Tesh, R.B., D.J. Gubler, and L. Rosen. 1976. Variation among geographic strains of Aedes albopictus in susceptibility to infection with chikungunya virus. Am. J. Trop. Med. Hyg. 25: 326-335. Theron, J., H. Huismans, and L.H. Nel. 1996a. Identification of a short domain within the non-structural protein NS2 of epizootic haemorrhagic disease virus that is important for single strand RNA-binding activity. J. Gen. Virol. 77: 129-137. Theron, J., H. Huismans, and L.H. Nel. 1996b. Site-specific mutations in the NS2 protein of epizootic hemorrhagic disease virus markedly affect the formation of cytoplasmic inclusion bodies. Arch. Virol. 141: 1143-1151. Tsetsarkin, K.A., D.L. Vanlandingham, C.E. McGee, and S. Higgs. 2007. A single mutation in Chikungunya virus affects vector specificity and epidemic potential. PLoS Pathog. 3: 1895-1906. Tsetsarkin, K.A. and S.C. Weaver. 2011. Sequential adaptive mutations enhance efficient vector switching by Chikungunya virus and its epidemic emergence. PLoS Pathog. 7: e1002412. Ulyshen, M.D. 2011. Arthropod vertical stratification in temperate deciduous forests: implications for conservation-oriented management. For. Ecol. Manage. 261: 1479- 1489. Uslu, U. and B. Dik. 2006. Vertical distribution of Culicoides larvae and pupae. Med. Vet. Entomol. 20: 350-352.

168

Uslu, U. and B. Dik. 2010. Chemical characteristics of breeding sites of Culicoides species (Diptera: Ceratopogonidae). Vet. Parasitol. 169: 178-184. Veggiani Aybar, C.A., M.J.D. Juri, M.S. Lizarralde de Grosso, and G.R. Spinelli. 2011. Spatial and temporal distribution of Culicoides insignis and Culicoides paraensis in the subtropical mountain forest of Tucumán, Northwestern Argentina. Florida Entomol. 94(4): 1018-1025. Venter, G.J., E. Hill, I.T.P. Pajor, and E.M. Nevill. 1991. The use of a membrane feeding technique to determine the infection rate of Culicoides imicola (Diptera: Ceratopogonidae) for 2 bluetongue virus serotypes in South Africa. Onderstepoort J. Vet. Res. 58: 5-9. Venter, G.J., J.T. Paweska, H. Lunt, P.S. Mellor, and S. Carpenter. 2005. An alternative method of blood-feeding Culicoides imicola and other haematophagous Culicoides species for vector competence studies. Vet. parasitol. 131: 331-335. Venter, G.J., P.S. Mellor, and J.T. Paweska. 2006. Oral susceptibility of South African stock-associated Culicoides species to bluetongue virus. Med. Vet. Entomol. 20: 329-334. Venter, G.J., I.M. Wright, T.C. Van Der Linde, and J.T. Paweska. 2009a. The oral susceptibility of South African field populations of Culicoides to African horse sicknes virus. Med. Vet. Entomol. 23: 367-378. Venter, G.J., K.G. Hermanides, S.N.B. Boikanyo, D.M. Majatladi, and L. Morey. 2009b. The effect of light trap height on the numbers of Culicoides midges collected under field conditions in South Africa. Vet. Parasitol. 166: 343-345. Venter, G.J., I.M. Wright, and J.T. Paweska. 2010. A comparison of the susceptibility of the biting midge Culicoides imicola to infection with recent and historical isolates of African horse sickness virus. Med. Vet. Entomol. 24: 324-328. Venter, G.J., S.N.B. Boikanyo, D.M. Majatladi, and L. Morey. 2016. Influence of carbon dioxide on numbers of Culicoides midges collected with suction light traps in South Africa. Med. Vet. Entomol. 30: 117-122. Veras, R.S. and E.G. Castellon. 1998. Culicoides Latreille (Diptera: Ceratopogonidae) in Brazilian Amazon. V. Efficiency of traps and baits and vertical stratification in the forest reserve Adolpho Ducke. Revta bras. Zool. 15: 145-152. Verdezoto, J., E. Breard, C. Viarouge, H. Quenault, P. Lucas, C. Sailleau, S. Zientara, D. Augot, and S. Zapata. 2018. Novel serotype of bluetongue virus in South America and first report of epizootic haemorrhagic disease virus in Ecuador. Transbound. Emerg. Dis. 65: 244-247.

169

Veronesi, E., G.J. Venter, K. Labuschagne, P.S. Mellor, and S. Carpenter. 2009. Life-history parameters of Culicoides (Avaritia) imicola Kieffer in the laboratory at different rearing temperatures. Vet Parasitol. 163: 370-373. Veronesi, E., M.R. Henstocks, S. Gubbins, C. Batten, R. Manley, J. Barber, B. Hoffmann, M. Beer, H. Attoui, P.P.C. Mertens, and S. Carpenter. 2013. Implicating Culicoides biting midges as vectors of Schmallenberg virus using semi- quantitative RT-PCR. PLoS One 8(3): e57747. Veronesi, E., A. Paslaru, C. Silaghi, K. Tobler, U. Glavinic, P. Torgerson, and A. Mathis. 2018. Experimental evaluation of infection, dissemination, and transmission rates for two West Nile virus strains in European Ades japonicus under a fluctuating temperature regime. Parasitol. Res. 117: 1925-1932. Vigil, S.L., J.C. Wlodkowski, and J. Parris. 2014. New records of biting midges of the genus Culicoides Latreille from the southeastern United States (Diptera: Ceratopogonidae). Insecta Mundi 0394: 1-14. Vigil, S.L., M.G. Ruder, D. Shaw, J. Wlodkowski, K. Garrett, M. Walter, and J.L. Corn. 2018. Apparent range expansion of Culicoides (Hoffmania) insignis (Diptera: Ceratopogonidae) in the southeastern United States. J. Med. Entomol. 55: 1043- 1046. Vose, J.M., B.D. Clinton, N.H. Sullivan, and P.V. Bolstad. 1995. Vertical leaf area distribution, light transmittance, and application of the Beer-Lambert law in four mature hardwood stands in the southern Appalachians. Can. J. For. Res. 25: 1036- 1043. Walker, E.D. and J.D. Edman. 1986. Influence of defensive behavior of eastern chipmunks and gray squirrels (Rodentia: Sciuridae) on feeding success of Aedes triseriatus (Diptera: Culicidae). J. Med. Entomol. 23(1): 1-10. Weaver S.C. and W.K. Reisen. 2010. Present and future arboviral threats. Antiviral Res. 85: 328-345. Weir, R.P., M.B. Harmsen, N.T. Hunt, S.D. Blacksell, R.A. Lunt, L.I. Pritchard, K.M. Newberry, A.D. Hyatt, A.R. Gould. and L.F. Melville. 1997. EHDV-1, a new Australian serotype of epizootic haemorrhagic disease virus isolated from sentinel cattle in the Northern Territory. Vet Microbiol. 58: 135-143. Wellby, M.P., M. Baylis, P. Rawlings, and P.S. Mellor. 1996. Effects of temperature on survival and rate of virogenesis of African horse sickness virus in Culicoides variipennis sonorensis epidemiology of the disease. Bull. Entomol. Res. 715-720. Wernike, K., B. Hoffmann, and M. Beer. 2015. Simultaneous detection of five notifiable viral diseases of cattle by single-tube multiplex real-time RT-PCR. J. Virol. Methods. 217: 28-35.

170

Wieser-Schimpft, L., W.C. Wilson, D.D. French, A. Baham, and L.D. Foil. 1993. Bluetongue virus in sheep and cattle and Culicoides variipennis and C. stellifer (Diptera: Ceratopogonidae) in Louisiana. J. Med. Entomol. 30: 719-724. Wiggins, K., B. Eastmond, and B.W. Alto. 2018. Transmission potential of Mayaro virus in Florida Aedes aegypti and Aees albopictus mosquitoes. Med. Vet. Entomol. 32: 436-442. Wilber, C.G. and P.F. Robinson. 1958. Aspects of blood chemistry in the white-tailed deer. J of Mammal. 39(2): 309-311. Wilson, W.C. 1991. Molecular comparison of VP3 from bluetongue and epizootic hemorrhagic disease viruses. Virus. Res. 21: 225-236. Wilson, W.C., E.S. O’Hearn, C. Tellgren-Roth, D.E. Stallknecht, D.G. Mead, and J.O. Mecham. 2009. Detection of all eight serotypes of epizootic hemorrhagic disease virus by real-time reverse transcription polymerase chain reaction. J. Vet. Diagn. Invest. 101: 97-101. Wisely, S.M. and K. Sayler. 2016. Autogenous vaccine field trial for epizootic hemorrhagic disease virus and bluetongue virus does not result in high titer to homologous virus serotypes. Int. J. Infect. Dis. 53S: 149-150. Wittmann, E.J., P.S. Mellor, and M. Baylis. 2002. Effect of temperature on the transmission of Orbiviruses by the biting midge Culicoides sonorensis. Med. Vet. Entomol. 16: 147-156. Wong, N.D., E.G. McDermott, A.C. Murillo, and B.A. Mullens. 2018. Field distribution and density of Culicoides sonorensis (Diptera: Ceratopogonidae) eggs in dairy wastewater habitats. J. Med. Entomol. 55: 392-397. Wood, W.F. 1997. Short-chain carboxylic acids in interdigital glands of gemsbok, Oryx gazella gazella. Biochem. System. Ecolog. 25(5): 469-470. Wood, W.F. 1998. Volatile compounds in interdigital glands of sable antelope and wildebeest. Biochem. System. Ecolog. 26(3): 367-369. Wood, W.F. 2003. Volatile components in metatarsal glands of sika deer, Cervus nippon. J. Chem. Ecolog. 29(12): 2729-2733. World Health Organization. 1967. Arboviruses and human disease: Report of a WHO Scientific Group. Worwa, G., M. Hilbe, V. Chaignat, M.A. Hofmann, C. Griot, F. EHrensperger, M.G. Doherr, and B. Thür. 2010. Virological and pathological findings in bluetongue virus serotype 8 infected sheep. Vet. Microbiol. 144: 264-273.

171

Zanella, G., B. Durand, E. Sellal, E. Breard, C. Sailleau, S. Zientara, C.A. Batten, P. Mathevet, and C. Audeval. 2012. Bluetongue virus serotype 8: Abortion and transplacental transmission in cattle in the Burgundy region, France, 2008-2009. Theriogenology 77: 65-72. Zhao L., B.W. Alto, D. Shin, and F. Yu. 2018. The effect of permethrin resistance on Aedes aegypti transcriptome following ingestion of Zika virus. Viruses 10: 1-18.

172

BIOGRAPHICAL SKETCH

Bethany McGregor grew up in Pontiac, Illinois later moving to Jackson,

Tennessee when she was eight. She graduated from the University of Tennessee–

Knoxville in 2012 with a B.S. in Wildlife and Fisheries Science with a focus on wildlife management. She went on to complete a M.S. in Biology at the University of Louisiana–

Monroe in 2015 studying the utility of inter-simple sequence repeats as genetic markers in Pleurocerid snails. She began her PhD studies in the fall of 2015 at the University of

Florida–Florida Medical Entomology Laboratory studying Culicoides midges on big game preserves where Culicoides-borne hemorrhagic diseases were causing morbidity and mortality to farm animals. Her research primarily focused on studying the ecology of

Culicoides and identifying the putative vector species of hemorrhagic disease in north

Florida. She graduated from the University of Florida in the spring of 2019 with a PhD in

Entomology and Nematology.

173