Defining Mechanisms Underlying Mite Tolerance and Honey Bee Survival

Progress report for GNE19-214

Project Type: Graduate Student
Funds awarded in 2019: $14,998.00
Projected End Date: 12/31/2021
Grant Recipient: The Pennsylvania State University
Region: Northeast
State: Pennsylvania
Graduate Student:
Faculty Advisor:
Expand All

Project Information

Project Objectives:

I will use two systems in which to examine how transmission rates influence selection of different genotypes of DWV (different strains) and virulence of these strains (as measured by mortality rates of infected bees). (1) Determining if rapid transmission selects for more virulent viral populations, by a) Generating populations of viruses that have been selected under conditions of rapid transmission in the laboratory, mimicking repeated transmission by Varroa b) Evaluating if these viruses exhibit increased levels of virulence in laboratory assays, relative to non-selected viral populations c) Determining if this variation in virulence is associated with underlying viral genetic variation by conducting whole-genome sequencing of the viral populations (2) Determining if viral populations in feral bees from the Arnot Forest have reduced virulence relative to managed honey bees, by a) Collecting viruses from feral bees and bees from nearby managed apiaries b) Sequencing viral populations to identify genetic differences associated with these isolated populations c) Assessing virulence of these viral populations Objectives (1a) and (1b) will be completed prior to the start of this proposal. Objectives (1c) and (2a-c) will be completed using the proposed budget, if funded. Results will be used to understand how viral populations can be managed to improve honey bee survival and adaptation. Furthermore, this work will be disseminated to researchers and beekeepers through publications and conferences, help develop outreach activities, and support undergraduate research experiences.

Introduction:

Approximately 75% of major food crops benefit from animal pollination (Klein et al., 2007), with pollination services contributing approximately $14-$23 billion annually to the US economy (Chopra, Bakshi, & Khanna, 2015). Declines in pollinators, and insect populations broadly, have been increasingly documented in recent decades (Hallmann et al., 2017; Kluser & Peduzzi, 2007; LeBuhn et al., 2012; Potts et al., 2010). Despite heavy management, US beekeepers lose an average of 40% of their colonies annually, with over 45% overwintering losses in Pennsylvania in 2017-2018 (Bee Informed Partnership, 2018). It is critical to develop sustainable approaches that mitigate the impacts of stressors to support food security and improve beekeepers’ economic outcomes.

The primary stressors driving honey bee colony losses in the Northeastern US and other temperate regions are Varroa destructor mites and Deformed Wing Virus (DWV) (Dainat, Evans, Chen, Gauthier, & Neumanna, 2012), which act synergistically to undermine bee health (Di Prisco et al., 2016). There are no methods to control DWV, though molecular approaches can be used to track levels and spread of viruses (Traynor et al., 2016). Beekeepers can only reduce DWV levels by controlling Varroa, which is accomplished primarily through repeated miticide application, which may also have negative effects on bees (Berry, Hood, Pietravalle, & Delaplane, 2013; Johnson, Pollock, Berenbaum, Anderson, & Varroa, 2009). Beekeepers typically purchase packages of bees from Southern states to replace lost colonies, but packages often have high Varroa loads and low survival rates, and thus is an expensive, unsustainable solution (SARE, 2014). As the cost of beekeeping increases, the cost of renting colonies for pollination services will also increase, which will lead growers to increase product prices to offset these costs (Calderone, 2012).

There is consequently considerable interest in breeding Varroa-tolerant or Varroa-resistant stocks of bees, including the PA Queen Improvement Project (NE SARE #FNE18-886). The ability to breed for resistance is supported by reports of feral bee populations which survive without management, such as the bees found in the Arnot Forest of New York (Seeley, 2007). However, traits associated with survival (small colony sizes, high queen turnover) are not sufficient to explain these high survival rates. I propose that the DWV population circulating within the Arnot Forest bees has been selected for reduced virulence due to decreased transmission rates, which have led to increased bee survival. If reduced transmission rate, leading to selection for decreased virulence (Schmid-Hempel, 2011), is a major determinant of feral honey bee population survival, then management practices must be developed to similarly decrease virus transmission rates in managed honey bee colonies.

Most efforts to develop sustainable approaches reducing the impact of honey bee stressors focus on breeding bees that are genetically tolerant or resistant to parasites or pathogens. Breeding bees which are resistant to Varroa mites typically focus on selecting for stocks that remove Varroa-infested pupae or groom Varroa off adult bees (Buchler, Berg, & Le Conte, 2010; Thomas, H, Gregory, & G, 2010). However, since honey bee queens mate with an average of 12 males far from their home colonies, it is difficult for beekeepers to maintain these traits in their stocks (Niño & Jasper, 2015). Wild populations in the Arnot Forest, though, are able to self-sustain and persist despite ubiquitous stressor exposure.

Varroa-mediated DWV transmission leads to increased titers and selection for more virulent strains (Di Prisco et al., 2016; Ryabov et al., 2014). High levels of DWV leads to deformed wings in adults, reduced activity and ability to contribute to colony tasks, and increased mortality (de Miranda & Genersch, 2010; McMahon et al., 2016; Nazzi et al., 2012). This increased mortality can unbalance the demographic structure of colonies and drive colony losses (Perry, Søvik, Myerscough, & Barron, 2016). Importantly, a vicious cycle is created by high levels of DWV that result in increased reproductive success of Varroa (Di Prisco et al., 2016), which then further transmit DWV to additional hosts. Thus, DWV, not Varroa, may be the more significant driver of colony losses.

How is it possible, then, that some feral bee populations are able to survive unmanaged despite the presence of Varroa and DWV? It is predicted that in populations where a virus cannot readily infect new hosts, where population size is small or hosts (i.e. colonies) are far apart (Nolan & Delaplane, 2017; Seeley & Smith, 2015), there should be selection for reduced viral virulence (Schmid-Hempel, 2011). These conditions may be met in the Arnot Forest, as colonies are small and more spread out than colonies in managed apiaries (Seeley, 2017). Conversely, easy transmission should select for more virulent strains, as predicted by Varroa-mediated vectoring (McMahon, Wilfert, Paxton, & Brown, 2018). I, therefore, hypothesize that the viral populations within these small, low-density feral populations have adapted reduced virulence due to decreased transmission rates.

To investigate my viral adaptation hypothesis, I propose to characterize the effects of viral transmission rate on DWV virulence and genotype, using viral populations generated from laboratory studies mimicking Varroa transmission (rapid pupae to pupae transmission), and viral populations collected from bees from the Arnot Forest and nearby, managed apiaries. These studies will both provide insights into how management practices can be adjusted to reduce selection for virulent genotypes and provide molecular markers to better screen for virulent viruses. This work will not only support honey bees and beekeepers, but additionally support vital wild bee pollinator populations, which are known to share pathogens with managed honey bees (Graystock, Blane, McFrederick, Goulson, & Hughes, 2016) and have also demonstrated population crashes in the past decades, promoting sustainable pollination and agriculture in the Northeast.

Research

Materials and methods:

Honey bee samples – Honey bees were in vitro reared (Schmehl et al 2016). This allowed for controlled development across samples, and also limited other unknown environmental variables that may affect DWV infection (including exposure to DWV from the colony). Bees were collected from a research colony with a single drone inseminated (SDI) queen – this allowed for approximately 75% relatedness between sister bees due to the honey bee’s haplodiploid sex-determination system, and therefore, minimized the effect of differing honey bee host genetics influencing DWV infection. This SDI colony was inspected weekly to ensure health status (ie low virus infection and low parasite load) and confirm the presence of the original queen. 

Virus Isolation – Since it is not possible to general purified bee viruses, crude isolations were conducted to create two starting inoculums, as well as the subsequent inoculums from the viral passaging paradigm (see below, Virus Passaging). Virus was isolated from individual bees using the following protocol. 2ml screw-cap microcentrifuge tubes with 2mm zirconia beads within were sterilized with a UV-cross linker for 2 minutes. Bees (flash-frozen and kept at -80°C until processing) were placed into 2ml screw-cap microcentrifuge tubes, and 500ul of 1x PBS was added. Bees were homogenized using a Bead Ruptor Elite (Omni International, Kennesaw, GA) at 6.5 ms for 45 seconds. Tubes were then placed on ice then centrifuged for 3 minutes at max speed. Supernatant was removed and passed through a sterile 0.2um syringe filter to separate viral particles from honey bee cells (pellets were stored at -80°C). This crude isolate was kept at 4°C for no longer than a week (for injections and/or RNA extractions) and stored long-term at -80°C.

Virus Passaging – The starting inoculum was injected into in vitro reared honey bee pupae at the white-eyed stage (14 days post-egg-laying). Virus concentration was approximately 6 x 10^9 genome equivalents. 1.5ul inoculum was injected using a mouth aspirator with an attached 10ul capillary tube pulled into a needle. Needles were changed between treatments to avoid contamination between treatments. Two separate starting inoculums were utilized. To measure colony DWV levels and the effect of the injection itself on DWV levels, control bees (in vitro reared but further unmanipulated) and PBS-injected bees (injected with the stock viral isolation saline) were included as controls (these groups are labeled as “no inject” and “PBS”). 

Two additional treatments were incorporated into the virus passaging paradigm to immunocompromise bees infected with DWV. Bees were either injected with a nylon thread to induce melanization (an immune response supplementary to the main antiviral response, the RNA interference pathway) or had 1μl hemolymph removed (to mimic Varroa feeding). These groups were identified as “DWV+mel” and “DWV+hemo”, respectively. Bees without any immunocompromisation were simply identified DWV+. Passaging virus through Varroa (the biological vector) was not included as proposed as mites could be successfully kept in the lab, but did not reproduce. To control for virus amplification triggered by the immunocompromisation protocol, bees were immunocompromised and injected with PBS (groups “DWV-mel” and “DWV-hemo”).

Virus inoculums from samples were isolated (see above, Virus Isolation) four days after infection, and all groups were passaged five times. See Figure 1 for all 10 included groups. 

Figure 1. Groups included in the immunocompromised bees viral passaging paradigm. 

To passage virus, viral inoculums with confirmed DWV (see below, Virus Quantification) was injected into a new in vitro reared honey bee pupae at the white-eyed stage using the described method – 1.5ul injection using a mouth aspirator and capillary needles – and isolated again 4 days post-injection. Inoculums isolated from bees with additional treatments (ie hemolymph removal or triggered melanization) were again injected into bees with the same treatment as the previous passage (ex DWV+hemo inoculum into another bee with 1μl hemolymph removed, etc.). This method was repeated until virus populations were passaged 5 times. 

Virus Quantification – RNA was extracted from 30ul of each virus inoculum using a Direct-zol RNA Miniprep kit (Zymo Research, Irvine, CA) following the manufacturer’s protocol. cDNA was prepared from each sample using a High-Capacity cDNA Reverse Transcription Kit with RNase Inhibitor (ThermoFisher Scientific, Maltham, MA) the following way : 1ul of 10x Buffer, 0.4ul of 25x dNTP mix, 1ul of 10x random primers, 0.5ul of Reverse Transcriptase, 0.5ul of RNAse Inhibitor, and 3.4ul of RNA. cDNA was diluted 1:20x to allow a sufficient amount of cDNA for all qPCR reactions. qPCR was conducted using PowerUp™ SYBR™ Green Master Mix (ThermoFisher) in the following way : 5ul SYBR, 1ul of 10x Forward Primer, 1ul of 10x Reverse primer, 1ul of water, and 2ul cDNA. qPCR was conducted using a 7900HT Fast Real-Time PCR system (Applied Biosystems) under the following parameters: 50°C for 2 minutes, 95°C for 10 minutes, then cycle 40x 95°C for 15 seconds and 60°C for one minute, and then a dissociation stage step for melting curve analysis. qPCR primers can be found in Table 1. 

Table 1. Primers used in this study to quantify DWV-A and DWV-B within viral populations. S8 was used to confirm successful RNA extraction and cDNA conversion. 

Primer name

Primer sequence (5’ to 3’)

Reference

DWV-A+B F

GTTTGTATGAGGTTATACTTCAAGGAG

Ryabov et al 2014

DWV-A+B R

GCCATGCAATCCTTCAGTACCAGC

DWV-B CP F

CTGTAGTTAAGCGGTTATTAGAA

Ryabov et al 2014

DWV-B CP R

GGTGCTTCTGGAACAGCGGAA

DWV-A CP F

CTGTAGTCAAGCGGTTACTTGAG

Ryabov et al 2014

DWV-A CP R

GGAGCTTCTGGAACGGCAGGT

DWV-B NS F

TTCATTAAAACCGCCAGGCTCT

Ryabov et al 2014

DWV-B NS R

CAAGTTCAGGTCTCATCCCTCT

DWV-A NS F

TTCATTAAAGCCACCTGGAACA

Ryabov et al 2014

DWV-A NS R

CAAGTTCGGGACGCATTCCACG

eIF3-S8 F

TGAGTGTCTGCTATGGATTGCAA

Alaux et al 2009

eIF3-S8 R

TCGCGGCTCGTGGTAAA

 

Arnot Forest Collections – Bees were collected from 13 sites across NY and PA (Figure 2). Collections were conducted between September 19 to October 14, 2019, between 10AM-5PM on sunny, warm (mid 60’s-mid 70’s°F) days, by Allyson Ray and Dr. Thomas Seeley. Bees were captured using insect nets and knocked out on dry ice and then placed into 15ml conical tubes by site and kept on dry ice. Wild Arnot Forest bees were collected while foraging on flowers (n = 5-12 bees/site), and bees from managed colonies were collected from colony entrances, preferentially selecting obvious foragers (indicated by pollen-filled corbiculae) when possible (n = 2-5 colonies/apiary, 10-15 bees/colony). Upon returning from the field, bees were placed at -80°C for long term storage. See Table 2 for details. 

Figure 2. Collection sites (From top to bottom, left to right : Site 3, 8, 9, 10, 11, 12, 13)

 

Table 2. 

Date of collection Site Location Colony # bees
Description / notes
9/19/2019 1 42.278775, -76.619141 N/A 10
Road on East boundary, abandoned barn
  2 42.261283, -76.628128 N/A 5
Sugar mill, south entrance to forest, small site, not much to collect 
  3 42.276844, -76.657628 N/A 11 near center? sites 3 and 4 were GREAT collecting sites
  4 42.285173, -76.658358 N/A 12
also near center, about 0.5mi as the bee flies 
  5 42.293468, -76.656313 N/A 6
North road, large logging site, dry and hard to collect from
  6 42.2944444,-76.6386111 N/A 11
Green Springs natural cemetery preserve
  7 42.256238, -76.656220 N/A 10
Southeast boundary, near road – bees on little white asters
  8 42.308280, -76.572549 1-5 15/colony
commercial apiary outside of Arnot
10/1/2019 9 42.434992, -76.390923 1 11
Ellis Hollow apiary – returning foragers
      2 10
Ellis Hollow apiary – Not super mean but defensive (Tom got stung!)
      3 11
Ellis Hollow apiary – no visible pollen on “foragers”
  10 42.386705, -76.394294 1 10
Brooketondale Apiary, Dunlap meadow – California Q
      2 10
Brooketondale Apiary, Dunlap meadow – Vermont Q
      3 10
Brooketondale Apiary, Dunlap meadow – Wild caught Q (captured summer 2018, so had overwintered)
  11 42.500471, -76.434303 1 10
Niemi Road Apiary – noticeable pollen collecting
      2 12
Niemi Road Apiary – lots of drones? Didnt see any pollen
      3 10
Niemi Road Apiary
10/9/2019 12 40.801257, -77.859579 1 12
Millennium Science Complex Rooftop apiary – Colony 50 (the “mean” one) 2 deeps and medium, strong, but lots of phoetic mites
      2 12
Millennium Science Complex Rooftop apiary – Colony 53, the one right next door, 1 deep 2 mediums, also strong – mite count 1 day prior : 6/300
10/14/2019 13 40.823654, -77.867992 1 12
Beneficial Arthropod Research Facility apiary – Colony 16, high mites
      2 12
Beneficial Arthropod Research Facility apiary – Colony 15 (12 mites/ 300 bees on 10/8/2019) Possibly new queen this 2019 Summer
      3 12
Beneficial Arthropod Research Facility apiary – Colony 36, low mites (2/300) w/ new queen this summer 2019
Research results and discussion:

Viral Passage Variant Quantification – Sample processing and quantification of viral variant populations are currently in progress – R0, R1, and R5 have been quantified (Figure 3), and quantification of R2-R4 were intended for Spring 2020, but the COVID-19 pandemic related shutdowns have delayed these experiments until Summer 2021, when it can be incorporated into undergraduate training and mentorship.  

Figure 3. Variant proportions from virus passaging experiments. 

Control samples did not have any detected virus (not shown). Intriguingly, while previous virus passaging experiments demonstrated universal trends in directional variant population shifts, much more variability occurred in this experiment. Both starting inoculums were almost entirely DWV-A, and while some experimental passages were nearly completely DWV-A after five passages, other lines accumulated DWV-B through the passaging paradigm. This happened across all DWV+ treatments, with no discernible pattern at this time. In the planned 2021 analyses, we will track the increase of DWV-B in these variant populations by quantifying R2-R4, and future work to assess single nucleotide variation within these populations may elucidate the mechanism behind the observed dynamics. 

Preliminary assessments of DWV infections in Arnot Forest bees and nearby apiaries – To assess whether DWV populations circulating within the Arnot Forest have been selected for reduced disease severity due to decreased transmission rates, DWV positive samples must be identified from each collection site. Sample processing and quantification are currently in progress, but a subset of samples across location types have been quantified. Overall, DWV levels across samples were low (Figure 4), which may be due to these bees being naturally infected, compared to experimentally infected as in our other experiments. I found the highest levels of DWV in individuals from the Arnot forest compared to nearby commercial apiaries (Arnot mean : 1.59E+06 ± 1.59E+06SE, Commercial mean : 2.73E+00 ± 1.60E+00). Surprisingly, the highest level of infection was identified in the nearby NY Research Apiaries (1.05E+07 ± 5.55E+06SE).  

Figure 4. Native DWV infection across locations in a subset of samples 

The higher levels of DWV in the feral bees of the Arnot forest compared to levels found in managed bees may suggest that DWV within the Arnot Forest has indeed evolved to be less deadly, allowing infected bees to remain healthy enough to carry infections into the foraging stage. This may increase the rate of transmission across colonies, as a potential trade-off to the reduced transmission within the small colonies. As this is an indirect and theorized assessment of virulence, survival assay from virus collected from these samples must be conducted in Summer 2021 to directly assess mortality differences caused by these different viral populations. As there was a range of infection levels, we hope to increase our number of DWV infected samples across locations by completing sample processing.   

Both DWV-A and DWV-B master variant strains were identified across samples, although not in the same pattern (Figure 5.) Some samples contained only DWV-B, such as samples from the Arnot forest, while other samples were only infected with DWV-A, or were co-infected with both variants, including samples from nearby research apiaries in NY. Interestingly, there is a wide range of infection levels across samples, and no variant is consistently outcompeting the other in co-infecting samples (i.e. DWV-B levels are not consistently higher than DWV-A levels, as observed by Norton et al 2020). After sample processing is complete, DWV-infected samples across locations will be submitted for next generation sequencing to identify any underlying DWV genetic variation. The bioinformatic analysis pipeline has already been developed for a  previous study (Ray et al., 2021 in prep), and will be efficiently adapted and applied to this project as well.  

Figure 5Proportion of DWV-A vs DWV-B across locations in a subset of samples 

Participation Summary
Any opinions, findings, conclusions, or recommendations expressed in this publication are those of the author(s) and do not necessarily reflect the view of the U.S. Department of Agriculture or SARE.