Abstract

The Dynamics and Emergence of Shared Viruses in Honey Bees and Native BeesBy Nina Sokolov Doctor of Philosophy in Integrative Biology University of California, Berkeley Professor Michael Boots, Chair

It is estimated that each year 40% of managed honey bee hives are lost in the UnitedStates. The little data we have on wild bees highlights significant declines in many species. Pathogens are implicated as one of the many important drivers behind losses. RNA viruses are of particular concern due to their high mutation rate and thus are considered the highest risk for spillover into a new species. Although these viruses are often described as ‘honey bee viruses’, their origins are almost entirely unknown, with many of these viruses also being found in wild bees. With the existing data we currently have, we cannot infer the direction of spillover, the main host reservoirs of the viruses, nor the ecological dynamics of these viruses. One way of addressing this issue is through longitudinal sampling in which different bee species are collected through time and their viruses quantified to make a time series. By monitoring these epidemiological dynamics, we can see if epidemics begin in certain species before being transmitted to others and infer transmission dynamics. These data are lacking but necessary to accurately assess the risk of viruses being transmitted between managed and wild bees. The severity of viral infections have increased after the recent emergence and spread of a novel parasitic mite, Varroa destructor. Varroa mites arrived in the United States only in the 1980’s and have subsequently made beekeeping a lot more challenging and unsustainable. This mite parasitizes honey bees by consuming their fat body, an organ essential for mounting immune responses, and acts as a vector for viruses. Before varroa mites most bee viruses experienced direct transmission where infectious virions were primarily being ingested by the bee host consuming contaminated floral products or interacting with conspecifics. With the arrival of varroa, this mite adds a novel transmission pathway in which the viruses are directly injected into the bee’s hemolymph, the circulatory fluid within invertebrates. This bypasses any of the viral defenses the bees have evolved in their gut lining; the virus passes directly into their blood and leads to higher viral loads in varroa-parasitized colonies. As a result, managed hives experience high overwintering losses. This has led to both increased use of pesticides against the mite (and subsequent evolution of resistance of the mite against the acaracides), and the pressure to breed honey bees that are resistant to varroa.

Notwithstanding these observations, studies of bee virus dynamics are lacking inCalifornia, even with a massive agricultural industry, and high native bee diversity. During the almond bloom in the Central Valley, over 1.7 million hives are shipped in on truck beds from across the country, potentially bringing novel pathogens to local sites. Subsequently, honey bees share floral resources at stopover sites with wild bees, which creates a prime zone for interspecies disease transmission. This mass mixing event has the potential to select for more deadly pathogens as these RNA viruses are greeted with an enormous influx of novel susceptible hosts to spread through. According to disease ecology theory, increased spatial connectivity between host populations allows for the evolution of increased virulence, meaning the deadliness of a pathogen. This could be acting as a mass mixing event that has the capacity to evolve more deadly diseases for bees, thus there is a pressing need to track the viral dynamics occurring in association with the almond bloom.

These pollination events in conjunction with contemporary beekeeping managementpractices all have a part to play in the disease ecology emerging within the honey bees. But pollinators do not exist in a closed system, and the consequences of industrial agriculture on honey bees are also being felt in the greater pollinator community. After managed honey bees participate in crop pollination events, they share floral resources with native bees at stopover sites between pollination events, creating a risk for inter-species transmission. As it stands, we know little about viral dynamics in honey bees at the population level, nor about the effect of crop pollination events on disease emergence in honey and native bees.

This dissertation is first comprised of a literature review on the status of efforts to breedhoney bees for varroa resistance but find that most of these actions are in fact breeding for parasite tolerance. I describe the distinct epidemiological, evolutionary and ecological consequences of tolerance for resistance to varroa. This is discussed for both the honey bees which are directly parasitized by varroa, and the indirect community level consequences of that parasitism on wild bees through shared viruses. During this time, I have developed three independent field systems in Northern California through a connection of local landowners, crop growers, gardeners, ranchers, and community members, along with commercial, professional and hobbyist beekeepers. This network allowed for me to consistently collect honey and native bees for viral quantification through time, resulting in a frozen time series collection of bees from 2019 – 2024. Two of these systems will be described in this dissertation while the third will be the focus of my postdoctoral research. These field sites allowed for the regular longitudinal sampling of bee viruses through time starting in February at the almonds blooming in the Central Valley. After the almonds finish blooming, this beekeeping operation moved their hives to the Sierra foothills for the spring before being experimentally placed at high elevation montane meadows in the Sierra for breeding projects and honey production. This time series data set allowed me to track honey bee viral dynamics in association with the almond bloom and then test the impact of commercially managed honey bees in a montane wildflower system.

Chapter 1: Avoiding the tragedies of parasite tolerance in Darwinian beekeeping.

Bee losses in the United States across the last several decades have been attributed, in part, to theparasitic virus vectoring mite Varroa destructor. In response, beekeepers have conducted breeding efforts to confer resistance to these invasive parasites. In this chapter I review the current literature on honey bee breeding for Varroa and highlight the many instances these bees are tolerating rather than resisting mite parasitization, and by proxy, tolerating viral infections. Although both strategies lead to greater survivorship amongst those tolerant honey bees, I stress the critical difference between resistance versus tolerance from an ecological and pathogen spillover perspective. Existing literature shows that selecting for tolerance will not only lead to more spillover from honey bees but may also select for viruses that are more virulent in wild bees leading to the ‘Tragedies of Tolerance’. I argue the evolutionary ecology of pathogen spillover must be considered in ecologically responsible honey bee management. By selecting for varroa resistance specifically we can improve both honey bee and wild bee health.

Chapter 2: Changes in the honey bee virome during the almond bloom in California

Natural migrations of animals as well as managed transport of agricultural species can enhancethe geographical spread of pathogens and increase interspecies transmission. According to disease theory, spatial structure can have impacts on the evolution of virulence, with increased connectivity leading to higher levels of pathogen virulence. In this chapter I track viruses in honey bees throughout the almond bloom to quantify viral dynamics through a mass pollination event. I sampled honey bee hives in February 2021 across four time points: before the almonds bloom, twice during the bloom and then at their next stopover site a month afterwards. Almond hives were compared to “control” hives that did not participate in the almond bloom and instead stayed at their home yard in Grass Valley, CA. I sampled 32 hives through time and quantified their viruses through RNA sequencing. Overall, I detected 34 viral species through time, which included RNA viruses, DNA viruses, phages, and plant viruses. Almond pollinating hives had higher viral species richness at all time points in comparison to the controls, except at the last time point when the almond pollinating hives returned to the control’s location. There it appears that the viral species richness was transferred to the control hives, with control hive’s viral richness significantly higher once the almond bees returned. Individual viral taxa experienced different temporal dynamics and association with the almond bloom. Black queen cell virus was consistently lower in almond pollinating bees, whereas the Lake Sinai Virus (LSV) complex was more prevalent in the almond hives. Apis mellifera filamentous virus showed viral dynamics that were independent of the bloom, and instead showed seasonal dynamics that were similar between locations. These results show that these mass migrations are impacting viral dynamics in honey bees, even in honey bees that did not participate in crop pollination. Additionally, certain viruses are more highly associated with the bloom than others which highlights focal viral species that should be monitored genetically. As these conditions are primed for the evolution of increased virulence, it is critical that we begin storing genetic data to act as possible repositories in the case of emergent new pathogen strains.

Chapter 3: Viral dynamics between managed honey bees and native wild bees.

Wildflower meadows in the Sierra Nevada Mountain range boast high levels of native bee andfloral diversity along with low honey bee abundance. This provides a unique environment to quantify bee viral dynamics in a field system that is isolated from agriculture. By taking advantage of the low honey bee density of this area, this chapter reports the results around a 'natural experiment,' using longitudinal sampling of wild bee populations for viruses before and after the human-mediated migration of commercial honey bees onto private land within the Tahoe National Forest. In this system we can ask whether viruses were in native bees before honey bee arrival, if honey bees bring viruses with them when they migrated, and whether there was evidence of spillover between species. Native bees including bumble bees (genus: Bombus), and mason bees (Genus: Osmia) were collected before and at three time points after honey bee arrival across eight sites in the summer of 2021. Honey bee hives that had previously participated in almond crop pollination were sampled before migration. Then, honey bee hive and foraging honey bees were collected at three time points after arrival in the Sierran meadows. Sites ranged in their abundance of honey bees foraging on flowers and in the presence of an apiary. The effect of honey bee abundance and apiary presence was tested as a possible explanation behind any patterns of viral prevalence in native bees. Deformed wing virus strain B (DWV-B), Sac brood virus (SBV), and Chronic bee paralysis virus (CBPV) were all already present in Bombus species before honey bee arrival, suggesting that bumble bees can independently maintain these viruses. Honey bees brought each of these viruses with them when they migrated. However, the temporal dynamics of each of these viruses were distinctly different. DWV-B increased in Bombus over time, regardless of honey bee abundance or apiary presence, indicating an epidemic within bumble bee populations. SBV was highly prevalent in Bombus early in the season and decreased over time in all species, while Osmia mason bees showed no SBV infection. Whereas honey bee foragers barely experienced any evidence of SBV, the hive samples exhibited high SBV prevalence which also declined through the season. As both Bombus and hive SBV prevalence decreased through time, this suggests a seasonality trend for this virus. CBPV exhibited high prevalence in honey bees but low prevalence in Bombus, suggesting honey bees as more competent hosts with minimal direct impact on bumble bees. Osmia mason bees had positive cases of both DWV-B and CBPV, but their prevalence levels were far lower than the honey bees and bumble bees. These insights emphasize the necessity for long-term temporal survey data to understand bee virus ecological dynamics and their potential impact on native bee health.