Here we describe two protocols: first to propagate, extract, purify, and quantify large quantities of honey bee non-enveloped virus particles, including a method for removing honey bee pupae and second to test the effects of viral infection using a highly repeatable, high-throughput cage bioassay.
Honey bees are of great ecological and agricultural importance around the world but are also subject to a variety of pressures that negatively affect bee health, including exposure to viral pathogens. Such viruses can cause a wide variety of devastating effects and can often be challenging to study due to multiple factors that make it difficult to separate the effects of experimental treatments from preexisting background infection. Here we present a method to mass produce large quantities of virus particles along with a high throughput bioassay to test viral infection and effects. Necessitated by the current lack of a continuous, virus-free honey bee cell line, viral particles are amplified in vivo using honey bee pupae, which are extracted from the hive in large volumes using minimally stressful methodology. These virus particles can then be used in honey bee cage bioassays to test inocula viability, as well as various other virus infection dynamics, including interactions with nutrition, pesticides, and other pathogens. A major advantage of using such particles is that it greatly reduces the chances of introducing unknown variables in subsequent experimentation when compared to current alternatives, such as infection via infected bee hemolymph or homogenate, though care should still be taken when sourcing the bees, to minimize background virus contamination. The cage assays are not a substitute for large-scale, field-realistic experiments testing virus infection effects at a colony level, but instead function as a method to establish baseline viral responses that, in combination with the semi-pure virus particles, can serve as important tools to examine various dimensions of honey bee-virus physiological interactions.
Honey bees (Apis mellifera) play a critical role in the modern global agricultural landscape but are currently suffering from a combination of biotic and abiotic stressors, including pesticide exposure, poor forage, parasites, and pathogens1,2. One of the most important pathogens of concern are viruses, many of which are vectored by another of the major honey bee stressors, the parasitic Varroa mite (Varroa destructor). These viruses can cause an array of negative effects in honey bees including reduced brood survival, developmental defects, and paralysis that can lead to total hive collapse both before and after overwintering periods3,4,5. Although there have been promising advances in the development of technologies used to combat virus infection6,7,8,9, the dynamics by which many viruses propagate, spread, and interact within a honey bee or colony are still poorly understood5,10. Understanding the basic biology of honey bee and virus interactions and their relationships with other environmental factors is critical for developing effective virus management techniques.
However, studying honey bee-virus interactions poses challenges with numerous known and unknown factors complicating the process. These include interactions with diet11,12, pesticide exposure13, and bee genetic background14,15. Even when focusing on virus infection alone, complications are common because honey bee populations, both managed and wild, always have some degree of background virus infection, though often without manifesting acute symptoms16,17, and the effects of virus coinfection are not well understood18. This has made the study of honey bee virus effects difficult to disentangle.
Many honey bee virus studies have used circumstantial virus infections to look for interactions with other stressors, observing how background infections change with other treatments12,19,20,21. While this approach has been successful at identifying important effects, especially discovering how pesticide or dietary treatments affect virus levels and replication, inoculation with a virus treatment of known content and concentration is critical for experimental testing of virus infection dynamics. Even so, separating experimental treatment from background infection can also pose challenges. In field studies, researchers have differentiated strains of deformed wing virus (DWV) to provide evidence for virus transmission from honey bees to bumble bees22, but using this approach would be difficult within honey bees alone. Virus infectious clones are a powerful tool, not just for tracking infection23,24,25 but for reverse genetics studies of honey bee viruses and for virus-host interaction research26,27,28. However, in most instances, infectious clones are still required to fulfill the infection cycle inside cells to produce particles. Such particles are preferred as inocula for experimental treatments because their infectivity is higher than the naked viral RNA and inoculation with encapsidated genomes mimics a natural infection.
The production of pure, uncontaminated honey bee virus inocula (wild-type virus strains or those derived from infectious clones) also pose challenges. These are primarily due to the difficulties in obtaining a reliable, continuously-replicating, virus-free honey bee cell line to produce pure-strain viruses29,30. While some cell lines have been produced, these systems remain imperfect; still, there is hope a viable cell line can be produced29, which would allow for finer control of virus production and investigation. Until such a line becomes widely available, most virus production protocols will continue to rely on the use of in vivo virus production and purification18,31,32,33,34. These approaches involve identifying and purifying virus particles of interest (or producing an infectious clone) and using them to infect honey bees, usually as pupae. The pupae are injected with the target virus and then sacrificed, and further particles are extracted and purified. However, because no bees are virus-free to begin with, there is always some degree of contamination from traces of other viruses in any such concentrate, and, therefore, great care must be taken in choosing bees with a low likelihood of background infections. Further, methods for removing the pupae from the comb cells for use in these protocols33 are very labor intensive and can induce stress in the bees, limiting production by these means18,32. Here, we report an alternative method that allows for large scale removal of larvae with little labor and less mechanical stress on the bees.
Once pupae are obtained and injected with the starting virus inoculum, they must be incubated to provide the virus time to replicate. Subsequently, produced virus particles can be processed into a form usable to infect experimental bees. There are several simple methods to achieve this, including using a crude homogenate35,36 or hemolymph generated from virally infected bees as a source of infection37. These methods are effective but run into a greater chance of introducing unknown variables from the background substrate (e.g., other factors in the dead bee homogenates). Additionally, it is desirable to concentrate the particles if an experiment requires giving a large, known dose of a virus in a short period of time. Therefore, for better control, it is preferable to use methods that allow for some level of purification and concentration of the virus particles. Generally, a series of precipitation and centrifugation steps will result in the removal of almost all possible non-target virus material33.
After producing this concentrated inoculum, it is beneficial to quantify the viral titers (qPCR) and characterize it with in vivo bioassays to test its viability and ability to cause mortality, as well as to corroborate that increased virus titers are obtained after infection. This can be achieved through injection experiments (either into pupae or adults) or feeding experiments (into larvae or adults). While all these approaches are possible, feeding to groups of adult bees in a cage is often the fastest and simplest. The cage assay method is also widely used for testing various other treatments on bees including pesticide toxicity38, ovary development39, and nutritional influence on behavior40,41 and, therefore, can form a good basis for experiments linking virus infection with other factors42.
Here we describe a reliable method for producing large quantities of semi-pure, highly-enriched virus particles without using an expensive ultracentrifuge, including a method for removing pupae that reduces labor and mechanical stress on the bees and a highly repeatable, high-throughput bioassay for testing viral infection and effects. By tightly controlling the purity of the viral inocula, investigators are able to reduce variation in honey bee viral response relative to other viral inoculation methods. Furthermore, the bioassay can screen for viral effects at a small groups level using highly repeatable experimental units before scaling to field-realistic settings, which is far more labor intensive to manage. In combination, these two methods provide the necessary tools for studies that can help improve our overall understanding of honey bee-virus physiological interactions.
1. Mass bee extraction option 1: larval self-removal
2. Mass bee extraction option 2: manual pupal excision
NOTE: Although option 2 (pupal excision) is a viable method of bee extraction, it also features several drawbacks when compared to option 1 (larval self-removal). Option 2 is far more labor intensive, harder to control for pupal age, and generally more stressful on the bees themselves. Option 1 is recommended whenever possible.
3. Pupal virus injection
NOTE: If performing this protocol for the first time (i.e., without prior viral inocula stocks), first extract and concentrate particles using adults, pupae, or larvae from a colony with a suspected infection. Measure the viral titers in the resultant inocula as described in step 5 and determine which particles to propagate further.
4. Virus particle concentration
NOTE: This protocol has not been tested for the recovery of enveloped viruses.
5. Virus RNA extraction and quantification
6. Viral feeding bioassay
Successfully following the protocols (Figures 1) for pupal injection and viral extraction should produce large quantities of virus particles. However, sampling and injecting pupae sourced from a variety of colonies at multiple time points maximizes the chances of acquiring target virus with low contamination. The dynamics by which viruses replicate and interact with one another within a honey bee is not well understood; coupled with the likelihood for preexisting infection, there is no guarantee that the injected (desired) virus will become the dominant in any given pupa, even if the pupae were sampled from the same original colony. Figure 4 demonstrates the potential range of viral proportions that one could expect to see following extraction. The four displayed colonies represent a subset of a larger virus harvesting effort in which every pupa was initially injected with a ~95% Israeli acute paralysis virus (IAPV) inoculum. Although 10 out of the 16 colony samples involved in these extractions contained highly pure IAPV (> 95%), including some > 99% (e.g., Colony 1), other samples varied in their IAPV proportion (e.g., Colony 2-3), with some even being dominated by other viruses such as deformed wing virus (DWV) (e.g., Colony 4).
Table 1 provides additional context for the amplification level of the four viruses (BQCV, DWV, IAPV, SBV) shown in Figure 4 in the form of RT-qPCR threshold cycle (Ct) values (the point at which a PCR target reaches the threshold of detection) and total virus genome equivalents (ge) per 100 ng RNA. Ct values can be used as a predictor of proportion, but ge values need to be calculated using a standard-curve based method17. Notice that the actual quantity of particles (i.e., ge) produced is dependent on the number of pupae processed and the filtering stringency during the extraction process.
Virus particle preparations (amplified inoculum) should be stored at -80 °C and it is recommended to aliquot them, as they will degrade significantly if subjected to multiple freeze-thaw cycles18. Additionally, dose-response assays should always be conducted prior to experimentation, as a multitude of external factors (hive genetics, bee health, etc.) can lead to highly variable viral response. Using data derived from experimental cages (Figure 3), Figure 5 demonstrates such variability by comparing the dose-response survival curves of honey bees, fed the same IAPV particles during two different years. Despite identical testing parameters, including the same viral inocula and testing concentrations ranging from 1% to 0.01% IAPV, the trials conducted in 2018 (Figure 5A) and 2019 (Figure 5B) produced noticeably different survival responses in all but the control treatment, which received no virus in its sucrose inoculum. Note that, if desired, LD50 calculations can also be performed at this point to obtain more precise mortality measurements43, but this is usually not necessary as approximations are generally sufficient. In 2018, a 1% dose resulted in approximately 50% survival at 72 hours post-infection (hpi), thereby making it the default concentration for most viral cage experiments conducted that year. However, that same dose achieved near total mortality in 2019 at the same timepoint, and as a result, most viral cage experiments conducted that year received a 0.01% IAPV inoculum instead. This significantly lower concentration reached the same levels of mortality as a 1% IAPV inoculum in 2018 while also using 100 times fewer virus particles.
These data were produced with bioassays using cages similar to the one diagrammed in Figure 3. The feeder holes in the top and side allow for easy diet control and the sliding cage door makes it simple to add or remove objects from the cage environment, such as inoculum trays or dead bees. However, the generalized viral assay protocol is not restricted to these types of cages nor diet choices and should be modified to suit experimental needs.
Figure 1: Representative images of various stages during the larval self-removal protocol. (A) Example larval mass that was expected following the overnight self-removal period (1.6). (B) Larvae spaced apart from one another on separate injection/growth trays (1.6.3) (C,D) White-eye pupae ready for viral injection (1.9). Brown spots are dried larval frass, which do not need to be removed. Please click here to view a larger version of this figure.
Figure 2: Example injector apparatus created by combining a hypodermic needle with a multi-dispenser tip. The needle and tip were joined using liquid adhesive to consistently deliver 1 µL fluid injections when attached to a repeating pipetter. Please click here to view a larger version of this figure.
Figure 3: Example cage used in virus bioassay. Sucrose solution and pollen were provided ad libitum though feeder holes during the duration of the trial. Virus inocula could be easily delivered using a tray inserted through the bottom of the cage. Note that the cage type and feeder content could be adjusted as necessary to suit experimental parameters. Please click here to view a larger version of this figure.
Figure 4: Average total virus loads and proportions across viral preparations from four sample colonies. Virus loads were measured by RT-qPCR as black queen cell virus (BQCV) + deformed wing virus (DWV) + Israeli acute paralysis virus (IAPV) + sacbrood virus (SBV) genome equivalents (ge) in 100 ng RNA. Each sample colony consisted of 150+ homogenized pupae originally injected with IAPV and represents the typical range of virus proportions generated from the pupal virus injection protocol. Please click here to view a larger version of this figure.
Figure 5: Dose-response survival curves of honey bee bioassay cages. Cages from both 2018 (A) and 2019 (B) were inoculated with IAPV and fed 30% sucrose solution ad libitum throughout the duration of the trial. Treatments represent IAPV inocula concentrations with 10-2 to 10-4 denoting 1% to 0.1% IAPV particle preparations mixed in 30% sucrose solution; control sucrose inoculations contained no virus. n = 56 total cages in 2018; n = 77 total cages in 2019. Please click here to view a larger version of this figure.
Table 1: Average threshold cycle (Ct) values and genome equivalents per 100 ng RNA of virus mixtures in four viral preparations from four sample colonies. Each sample colony consists of 150+ homogenized pupae generated from the pupal virus injection protocol. Viruses were detected by RT-qPCR using specific primers for each virus. Please click here to download this table.
Here we have outlined methods detailing every step of the virus amplification and inoculum stock preparation process, including larvae collection and virus propagation, extraction, and concentration, as well as viral treatment in the form of cage-feeding experiments. These methods allow for production of semi-pure virus particles (Figure 4), the effectiveness of which can be consistently be quantified by dose-response mortality testing for viruses that are lethal to adults (Figure 5). Following confirmation of infective ability and/or pathology, the generated particles can then be used in bioassays to elucidate the interactions between honey bees and honey bee viruses.
One of the most distinctive benefits of the described protocols is that each is easily scaled to a large volume, whether it be pupae harvested, particles produced, or bioassays performed. Using pupal excision methods33, one can expect to remove ~100 pupae per hour, though this number will scale with proficiency. However, the larval self-removal method, reported here, while requiring different planning and scheduling procedures, can easily generate 10-20 times that number of removed bees overnight while involving comparatively little manual effort and less mechanical strain on the bees.
The main limitation of this method is that there is no way of guaranteeing that the self-removed larvae were not previously Varroa-infested. Regular mite treatment and monitoring of source hives can minimize this risk, but some larvae still may have had some level of Varroa parasitization. Manually removing pupae individually, however, allows the user to observe if any mites are present in the cell of a given pupa. Additionally, because the self-removal process relies on the food-seeking behavior of the larvae, the frames containing these larvae must be removed before the final feeding that normally occurs. Only due to this lack of provisioning by the workers do the larvae crawl from their cells. Therefore, the pupae derived from this method experience a very short window of nutritional stress compared to larvae that developed completely inside a colony and can appear slightly smaller. This is particularly notable in the youngest larvae in the cohort present on the frame; because the queen has laid eggs over a 24 h window, these larvae are missing proportionally more feeding time. These are usually clearly notable during pupation for their very small size compared to those removed by manual excision. However, the volume of larvae produced by self-removal more than compensates for their diminished individual biomass. Furthermore, the manual excision method also can cause substantial mechanical stress to the pupae as they are pulled from their cells. If either method is to be used as part of an experiment, and not just to produce virus particles, care should be taken to ensure proper controls.
Regardless of the extraction method, the virus propagation protocol can generate large quantities of virus particles using the pupae, which minimizes the variability induced by other currently practiced virus inoculation methods35,36,37 when testing for honey bee viral response. It is important to note that this protocol was optimized and tested using non-enveloped viruses in the order Picornavirales (e.g., Israeli acute paralysis virus, deformed wing virus). Different strategies to isolate viral particles should be followed when working with enveloped viruses44. As a rough approximation, each of the 16 samples involved in the virus harvesting effort (which included the 4 colonies of Figure 4) were generated from 200-300 injected pupae and yielded between 2-2.5 mL of concentrated virus particle preparations. Assuming a virus inoculum concentration of 0.1% and a standard 35-bee cage, each of the 16 virus preparations would provide sufficient particles for 3,300-4,200 cages. This surplus of infective material reduces experimental restrictions and enables high-throughput bioassays. It is important to note that although the virus particle concentrate can remain viable for months or years when stored at -80 °C, it can slowly decrease in infectivity, even if subjected to few freeze-thaw cycles. It is, therefore, recommended to store the viral preparation stock in small aliquots, several of which can then be used to quantify viral titers and verify the treatment dose at the time of the experiment. Additionally, the inter-year variability demonstrated in Figure 5, further reinforces the need for periodic dose-response testing, thereby minimizing the chances of unexpected loss in virus viability.
The cage bioassays themselves also have limitations, or at least caveats necessary to take into consideration, the primary one being the artificial nature of the cage bioassay environment (Figure 3). Grouping bees into enclosures allows for viral testing beyond the individual level; the cage becomes the experimental unit rather than the bee itself. Although this is more similar to an actual colony than a single bee being treated in isolation, it is still far from a realistic hive environment. Removed from the social environment, including queen and brood pheromones, bees of different ages, and other cues, these bees may not respond as a full-sized colony might. These are primarily considerations for larger-scale experiments using the cage system, however. The results gathered from cage viral bioassays should primarily be treated as baseline information establishment that can be used to inform future virus testing decisions scaled to a more field-realistic setting as desired by the user.
The methods described here provide a standardized process for the mass production of virus particles for use in honey bee viral assays. Such assays have already been implemented to examine various aspects of honey bee-virus interactions, including multi-virus infection and how diet quality and nutritional supplements affect survivorship in the face of viral infection11,18,45,46. They have been scaled up for use in colony-wide infection experiments11,47 and to study the effects of infection on behavior47 and gene expression48. Overall, these methods provide a baseline in tools that can be used to produce and evaluate honey bee virus inocula.
The authors have nothing to disclose.
We would like to thank Dr. Julia Fine for her ideas and discussion during the protocol creation process, as well as Dr. Cassondra Vernier for her helpful comments throughout editing. These materials contributed towards projects that were supported in part by the Foundation for Food and Agriculture Research, under grant ID 549025.
10% bleach solution | |||
24:1 chloroform:isoamyl alcohol | SigmaAldrich | C0549 | |
70% ethanol solution | |||
Cages for bioassay | Dependent on experimental setup | ||
Combitips Advanced 0.1 mL | Eppendorf | 30089405 | Optional (if no injector appartus is available) |
Containers for larval self-removal | Should measure roughly 19" x 9-1/8" (Langstroth deep frame dimensions) | ||
Forceps | Blunt, soft forceps for larval separationl; blunt, hard forceps for pupal excision | ||
Fume hood | |||
Incubator | Capable of maintaining 34 ºC and 50% relative humidity | ||
Kimwipes | Fisher Scientific | 06-666 | Any absorbent wipe will work |
Medium-sized weight boats | Serve as inoculum trays | ||
Microcon-100kDa with Biomax membrane | MilliporeSigma | MPE100025 | |
NaCl | |||
Nitrile gloves | |||
Phosphate buffered saline (PBS) | SigmaAldrich | P5119 | |
Polyethylene glycol 8000 (PEG) | SigmaAldrich | 1546605 | |
Refrigerated benchtop centrifuge | Capable of 15,000 x g | ||
Refrigerated centrifuge | Capable of 21,000 x g | ||
Repeater M4 Multipipette | Eppendorf | 4982000322 | Optional (if no injector appartus is available) |
RNAse Away | ThermoFisher | 7000TS1 | |
RNAse-free water | SigmaAldrich | W4502 | |
Sucrose | |||
TES | SigmaAldrich | T1375 |