General enquiries on this form should be made to:Defra, Science Directorate, Management Support and Finance Team,Telephone No. 020 7238 1612E-mail: research.competitions@defra.gsi.gov.uk

SID 5 Research Project Final Report

SID 5 (2/05) Page 1 of 33

NoteIn line with the Freedom of Information Act 2000, Defra aims to place the results of its completed research projects in the public domain wherever possible. The SID 5 (Research Project Final Report) is designed to capture the information on the results and outputs of Defra-funded research in a format that is easily publishable through the Defra website. A SID 5 must be completed for all projects.

A SID 5A form must be completed where a project is paid on a monthly basis or against quarterly invoices. No SID 5A is required where payments are made at milestone points. When a SID 5A is required, no SID 5 form will be accepted without the accompanying SID 5A.

This form is in Word format and the boxes may be expanded or reduced, as appropriate.

ACCESS TO INFORMATIONThe information collected on this form will be stored electronically and may be sent to any part of Defra, or to individual researchers or organisations outside Defra for the purposes of reviewing the project. Defra may also disclose the information to any outside organisation acting as an agent authorised by Defra to process final research reports on its behalf. Defra intends to publish this form on its website, unless there are strong reasons not to, which fully comply with exemptions under the Environmental Information Regulations or the Freedom of Information Act 2000.Defra may be required to release information, including personal data and commercial information, on request under the Environmental Information Regulations or the Freedom of Information Act 2000. However, Defra will not permit any unwarranted breach of confidentiality or act in contravention of its obligations under the Data Protection Act 1998. Defra or its appointed agents may use the name, address or other details on your form to contact you in connection with occasional customer research aimed at improving the processes through which Defra works with its contractors.

Project identification

1. Defra Project code HH0819SHB

2. Project title

Fungal control of Varroa destructor

3. Contractororganisation(s)

    University of WarwickWarwick HRIWellesbourneWarwickCV35 9EF UK     

54. Total Defra project costs £ 383,802

5. Project: start date................ 01 September 2002

end date................. 31 March 2006

SID 5 (2/05) Page 2 of 33

6. It is Defra’s intention to publish this form. Please confirm your agreement to do so...................................................................................YES NO (a) When preparing SID 5s contractors should bear in mind that Defra intends that they be made public. They

should be written in a clear and concise manner and represent a full account of the research project which someone not closely associated with the project can follow.Defra recognises that in a small minority of cases there may be information, such as intellectual property or commercially confidential data, used in or generated by the research project, which should not be disclosed. In these cases, such information should be detailed in a separate annex (not to be published) so that the SID 5 can be placed in the public domain. Where it is impossible to complete the Final Report without including references to any sensitive or confidential data, the information should be included and section (b) completed. NB: only in exceptional circumstances will Defra expect contractors to give a "No" answer.In all cases, reasons for withholding information must be fully in line with exemptions under the Environmental Information Regulations or the Freedom of Information Act 2000.

(b) If you have answered NO, please explain why the Final report should not be released into public domain

Executive Summary7. The executive summary must not exceed 2 sides in total of A4 and should be understandable to the

intelligent non-scientist. It should cover the main objectives, methods and findings of the research, together with any other significant events and options for new work.

The varroa mite, Varroa destructor is an invasive and highly damaging ectoparasite of the European honey bee Apis mellifera . Beekeepers are attempting to control varroa with conventional pesticides, but resistance to these chemicals is developing and alternative, sustainable methods of control are required urgently. Previous Defra-funded research by our project team showed that varroa was susceptible to infection by entomopathogenic fungi. The purpose of the present study was to investigate these fungi as biocontrol agents of varroa.

Objective 1: conidia production and survival. Using data from previous research, 10 fungal isolates were selected from 6 species to take forward for more detailed examination against varroa: for this report the species are described as A - F. Five of the isolates are used in commercial products. Single spore isolates (SSIs) were prepared and placed into cryo-preservation. A two stage mass production system was developed which gave acceptable yields of conidia for experimental purposes, although higher yields would be needed for commercial production. The viability of conidia over 48 weeks was improved by storing them at low temperature, but none of the isolates showed evidence of being stable at room temperature for the 12 – 18 month target period that would be required for a commercial product. It is likely that the shelf life of conidia could be improved considerably by advances in processing and storage and this will be an important area for future work. Laboratory experiments indicated that conidia of species A are likely to persist within a bee hive for considerably longer than other fungal species.

Objective 2: Activity against varroa mites of fungi applied to bees. In previous research, bioassays were done that enabled fungal isolates to be identified that did not infect A. mellifera by conventional contact action. However, it is also possible that conidia applied to a hive could be ingested by bees and cause deleterious effects. A ‘worse case’, no-choice bioassay was done in which adult bees were fed a liquid suspension of fungal conidia. Species E & F were non pathogenic whereas isolates of species A and C caused approx. 80% mortality after 15 days. A more likely way in which adult bees could ingest fungal conidia is via consumption of contaminated pollen. When adult bees were fed a mixture of pollen and conidia, it was found that species C isolates were repulsive to bees. However, on the occasions when they were consumed, significant mortality was caused only by one species C isolate. Isolates of species A, E & F caused zero to low levels of mortality. Honey bee larvae may differ in their susceptibility to entomopathogenic fungi compared to adult bees, and the routes of exposure to fungal inoculum are likely to be different. Experiments were done in which individual honey bee larvae on brood comb were fed conidia suspensions of species C & F administered to brood food.

SID 5 (2/05) Page 3 of 33

While species F had no effect on larvae, larvae fed species C were less likely to survive. Moreover the fungus was transferred to adjacent control larvae by nurse bees, causing mortality. However, larvae that survived to capping pupated as normal.

Research was also done to provide data on fungal efficacy against varroa in different spatial scales. In a laboratory experiment, it was found that mites, placed on bees treated earlier with a technical powder of species F became infected and died (average LT50 = 6.75 days). The effect of the technical powder on varroa mites in a nucleus bee colony was investigated in a flight room experiment. Following application of the fungus (done on two separate occasions, either by inoculating adult bees or by applying powder to the top of frames), there were high levels of mycosis in dead mites that dropped from the colony. Mycosed mites were evident for approx. 13 days after each treatment. There was considerable natural variation in the daily mite drop prior to the first treatment, which was also high relative to later in the experiment. This may have been a result of disturbance when the colony was established. It will be important in future experiments to account for this underlying variation in future experiments. Nevertheless, the results indicate that the fungus caused high levels of infection in the varroa population. Priorities for future research will be to refine the method used to assess fungal efficacy and to develop an efficient application system, including a method to prolong the persistence of fungal induced mortality in varroa.

Objective 3: Interaction of fungal conidia with bee products and the environment. Laboratory experiments to measure the survival of fungal conidia in honey showed that conidia survived for no longer than 7 days at 20C. Survival varied with fungal isolate, and conidia of species C died within hours of being held in honey. Nucleotide sequence information of the rRNA gene repeat unit was obtained for isolates of species A, B, C, & F. Forward and reverse primers were designed for each fungus and evaluated for their specificity. The movement of species F commercial technical powder within a bee colony and onto flowers being visited by treated bees was investigated in the flight room experiment described above using nested PCR. The results indicated that species F is present naturally at low levels on bees. Post treatment, species F was detected in all samples of flowers, foraging worker bees and brood nest adult workers. The fungus was also detected in pollen cells, but despite being applied directly into the colony, it was detected in only 20% of honey cells. It will be important in future to design a fungus application system that minimises the probability of conidia entering honey cells

Objective 4: Engagement with key actors. Veterinary medicine products (which include pesticides applied to bees) are regulated in the UK by the Veterinary Medicines Directorate (VMD). The VMD declined our request for an overarching discussion on how biological control agents could fit in with the veterinary medicines registration system. Instead, the regulator appeared to be focused on specific details relating to dossier submission. There is a risk that regulating biocontrol agents according to a model designed for chemicals will act as a barrier to commercialisation, and is indicative of the bureaucratic tendency for policy instruments to be considered in isolation from their wider effects. There are important parallels to be drawn here with systemic failures in the regulatory system used for microbial biopesticides as plant protection products and the development of innovative regulatory measures by the Pesticides Safety Directorate.

Given the spread of chemical pesticide resistance in varroa, it is essential that new methods for its control are developed. Our research indicates that entomopathogenic fungi have potential as biological control agents of varroa. Although more research is required, the costs of developing a fungal biological agent are likely to be markedly lower than the costs of developing a new chemical pesticide. A fungal control agent is unlikely to be a ‘magic bullet’ solution for varroa, but it could be a valuable control option as part of a sustainable, integrated pest management system. The systems and approaches developed in this project could also be used to develop other alternative biologically-based agents, should this be required. A central theme of this research has been to identify and overcome scientific and technical barriers to the development of new, sustainable methods of pest control. However, in order to make sustainable varroa management a reality, it will be necessary to overcome regulatory barriers as well.

SID 5 (2/05) Page 4 of 33

Project Report to Defra8. As a guide this report should be no longer than 20 sides of A4. This report is to provide Defra with

details of the outputs of the research project for internal purposes; to meet the terms of the contract; and to allow Defra to publish details of the outputs to meet Environmental Information Regulation or Freedom of Information obligations. This short report to Defra does not preclude contractors from also seeking to publish a full, formal scientific report/paper in an appropriate scientific or other journal/publication. Indeed, Defra actively encourages such publications as part of the contract terms. The report to Defra should include: the scientific objectives as set out in the contract; the extent to which the objectives set out in the contract have been met; details of methods used and the results obtained, including statistical analysis (if appropriate); a discussion of the results and their reliability; the main implications of the findings; possible future work; and any action resulting from the research (e.g. IP, Knowledge Transfer).

tron

1. Introduction The varroa mite, Varroa destructor is a highly damaging ectoparasite of the European honey bee Apis mellifera. This pest originates in Asia, where it is a parasite of the Eastern honeybee, Apis cerana, but it has extended its host range to A. mellifera and has caused severe damage to populations of this species world-wide in recent years (Ball, 1994a; Oldroyd, 1999). It entered Europe in the 1970s, the USA in 1987 and the UK in 1992 (De Jong et al., 1982; Delfinado-Baker & Houck, 1989; Walton, 1996; Martin, 1997). At present, beekeepers attempt to control varroa with chemical pesticides, but resistance to these chemicals is developing (Thomas, 1997; Elzen et al., 1998) and alternative, sustainable methods of control are required urgently.

Varroa mites feed on the haemolymph of honey bee pupae and adults, and in so doing they transmit diseases that reduce bee longevity, lower reproductive capacity, and induce deformities (Ball, 1993, 1994a, 1994b; Martin, 1997; Bowen-Walker et al., 1999). The lifecycle is haplodiploid, and reproduction occurs in the brood cells of the colony. Details of the lifecycle are given by De Jong et al. (1982), Ramirez (1987), Beetsma (1994), Donze and Guerin (1994) and Sammataro et al. (2000). Mated female mites enter brood cells to oviposit shortly before the cells are capped. The first egg to be laid is unfertilised and develops into a neotenic male that remains in the cell. Subsequent eggs are fertilised and develop as females, which mate with their brother soon after maturation. The developing mites feed from the haemolymph of the bee larva / pupa through a single hole constructed and tended by the parent mite. Adult females emerge with the young bee 12 – 14 days after capping. On A. cerana, the reproduction of V. jacobsoni is restricted to drone brood and does little harm. Populations of V. destructor n. sp. that affect A. mellifera, however, also reproduce on worker brood, which is extremely debilitating to the colony (Oldroyd, 1999). Reductions in the numbers of worker bees leads to poor brood care, reduced colony homeostasis, and a concomitant increase in brood diseases (Ball, 1994b).

Apis mellifera is utilised widely for crop pollination and losses due to varroa can impact significantly on agriculture, horticulture and wild plants. The contribution of honeybees to society are difficult to quantify, but in the UK their activities have been valued at £120 million (ADAS Defra report, need reference). In the USA, the value is estimated at £12.5 billion ($19 billion) per annum (Beetsma, 1994). The ecological and economic impact of varroa is considerable. The winter kill of managed honeybee colonies by varroa was estimated at 13 million colonies world-wide in 1996 (Sanford, 1996), equivalent to a quarter of the global commercial population. Colony losses of up to 65% have been reported in some countries (Matheson, 1994). Losses due to varroa show no signs of reducing over time. For example, it is reported to have caused a loss of 50% of the honey bee population in the USA in 2005. In addition, varroa has destroyed nearly all stocks of feral bees (Martin, 1997) and is considered a threat to biodiversity (Allen-Wardell et al., 1998).

At present, varroa is controlled with pesticides, but resistance to these chemicals has developed and is spreading (Milani, 1994, 1999; Hillesheim et al., 1996; Thomas, 1997; Elzen et al., 1998). The USA (with the exception of Florida) and UK currently permit only pyrethroids for varroa control (Bew, 1992; Ball, 1994a; Sanford, 1999) which could increase the selection pressure for resistance in these countries. Pyrethroids and organophosphorus pesticides, sprayed against varroa, have a propensity to accumulate in beeswax (Fries, 1997). Alternative agents, such as organic acids and essential oils, and cultural control methods (heating, drone trapping), are only partially effective and are labour intensive (Fries, 1993; Mobus & De Bruyn, 1993; Ritter, 1993; Ball,1994a; Beetsma, 1994; Engels, 1994). The breeding of varroa-tolerant bees is a longer term option (Beetsma, 1994; Buchler, 1994; Danka et al., 1995; Harbo & Hoopingarner, 1997). There is a risk,

SID 5 (2/05) Page 5 of 33

however, that any varroa tolerance bred into commercial bee stocks could be dissipated by interbreeding with feral bees.

Alternative control agents for varroa are required urgently, therefore. Previous DEFRA-funded research identified entomopathogenic fungi as having potential for varroa control (HH0811SHB; see Chandler et al., 2001) and adult varroa mites were found to be highly susceptible to infection by entomopathogenic fungi in laboratory bioassays (HH0813SHB; see Shaw et al., 2002). The purpose of this study was to extend these investigations with the longer term goal of implementing practical cost-effective and sustainable biocontrol of varroa. The study also aims to advance our scientific knowledge of fungal ecophysiology and the behaviour of social insects, and the interactions between trophic levels.

2. Project Objectives The overall aims of this study were to: (a) investigate the use of entomopathogenic fungi as microbial control agents of the varroa mite; (b) quantify tritrophic interactions in a unique system of social insect - arachnid parasite - fungal pathogen. The component objectives were as follows:1. Fungus optimisation. Determine the effects of abiotic factors on conidia survival and production.2. Effect of fungus in varroa and bee populations. Quantify the activity against varroa mites of fungi

applied to bees. 3. Environmental impact. Investigate the potential for fungal conidia to be disseminated to bee products

and the environment. 4. Knowledge transfer. Engage stakeholders and exchange information.

3. Scientific and technical progress3.1. (Objective 1): Determination of the effects of abiotic factors on conidia production and survival.

3.1.1 Isolate selection.In previous DEFRA funded research (HH0813SHB), we examined 40 isolates of entomopathogenic fungi from six genera of the anamorphic Ascomycetes against varroa in laboratory bioassays. For this study, we selected 10 of these isolates from four genera to take forward for more detailed examination against varroa (Table 1). Using data from HH0813SHB, the isolates were selected on the basis of virulence to varroa mites, response to the environmental conditions likely to be experienced in bee colonies, and effects on non-target organisms measured in laboratory bioassays. Five of these isolates caused 100% mortality of varroa mites within seven days under simulated bee colony conditions (30ºC; 40% RH) in the laboratory (isolate codes A1, C4, C5, C6, E9). Five of the isolates are also used in commercial products. Most of the work in the current project focuses on A1, B3, C4, and F10.

The fungal isolates were stored as cryo-preserved stocks of conidia (Chandler, 1994), from which cultures were grown on Sabouraud dextrose agar slopes (SDA; 4% D-glucose, 1% mycological peptone, 2% technical agar No.3, prepared in de-ionised water) and maintained at 4ºC for up to six months. For experiments, cultures were grown from the slopes on SDA Petri plates for 14 d at 23C (species A,C,D,E,F) or 27C (species B), then conidia harvested in sterilised 0.03% Tween 80, filtered through sterile muslin to remove hyphal fragments and enumerated using an improved Neubauer haemacytometer. Viability was assessed by measuring the germination of conidia on SDA after incubation for 24 h at 23ºC (Goettel and Inglis, 1997). Viability was never less than 91% unless stated otherwise. Experiments that investigated the storage longevity of spore powders and their effects on honeybees used aerial conidia that had been mass produced on rice (see 3.1.2 below). The nucleus colony experiment (3.2.4) used a technical spore powder of isolate F10 supplied by a commercial company.

Production of single spore isolates and bioassays. Experiments to develop a molecular biology method for the detection of fungal isolates in environmental samples (3.3.1. and 3.3.3) required the production of single spore isolates (SSIs). Single spore isolates are also attractive for industry because they make molecular detection of the isolate more certain, which can be important for quality assurance or protecting intellectual property. Single spore isolates were prepared for all 10 isolates as follows: Aliquots of conidia suspensions (20 μl, 1 x 103 conidia ml-1) prepared from 10 d old SDA cultures were spread onto thin SDA plates and incubated for 24 h at 23C. Plates were then inspected using a Nikon SM2-2T stereo microscope, and individual, germinating conidia were cut out using a scalpel and transferred to separate SDA plates. Cultures were grown at 23C until plates were fully colonised, after which spores were harvested and placed into cryopreservation. Between 7 to 30 single spore isolates were obtained for each isolate and all have been placed in the Warwick HRI culture collection (134 in total).

SID 5 (2/05) Page 6 of 33

Laboratory bioassays were then done to evaluate the virulence to V. destructor of 10 SSIs (Table 2). General sources of A. mellifera and V. destructor are given in 3.2.1. Honey bee brood frames were infested with V. destructor and placed inside four honeybee colonies. Population levels of V. destructor were monitored weekly. Adult female V. destructor for laboratory bioassays were collected by hand from sealed worker brood cells. White-eyed bee pupae were collected by hand from uninfested frames as a food source for the mites. Mites were held overnight on bee pupae in 1.5 ml microfuge tubes plugged with dental roll at 30°C and 100% RH in darkness before treatment. Five V. destructor were placed on each bee pupa. Ten single spore isolates of entomopathogenic fungi were tested in this study (Table 1).

Conidia suspensions were prepared at concentrations of 106 conidia ml-1(species B) and 108 conidia ml-1 (all other isolates) in sterile 0.03% Tween 80 as described previously. Suspensions were held on ice for a maximum of 24 h before assay. Viability was assessed by measuring the germination of conidia on SDA after incubation for 24 h at 23ºC (Goettel and Inglis, 1997). Groups of five V. destructor were immersed for 10 s in 5 ml conidia suspension or 0.03% Tween 80 as a control. Excess suspension was removed by filtration under vacuum through filter paper (Whatman No.1), and the mites plus the wetted filter paper were transferred to a 9cm diameter Petri dish. This was sealed with Parafilm and kept on the laboratory bench for 1 h. Each group of five mites was then transferred to a 1.5 ml microfuge tube containing a fresh white-eyed honeybee pupa. Each fungal isolate was evaluated against two batches of five mites (four batches for the control). The microfuge tubes were placed horizontally on racks inside clear polypropylene containers (293 x 202 x 130 mm) with ventilated push-fit lids. Containers were maintained in the dark at 25°C and 100% RH (humidity obtained by placing 1% water agar in two 9cm Petri dish bases in the bottom of the container). The temperature and humidity within the containers were monitored using Squirrel data loggers (Grant Instruments, Cambridge, UK). Bee pupae were replaced every 6 d. Dead mites (no movement or response to stimulus) were removed and incubated on 1% water agar at 23°C. The presence of sporulating mycelia on mite cadavers was used as an indication of fungus-induced mortality.

Data for the virulence of the SSIs is given in Table 2. There were not sufficient resources to bioassay the SSIs and their parent cultures together in this project. However, in general the mean times to death of the SSIs were in keeping with those measured previously for the parent cultures (Shaw et al., 2002) which ranged from 44 h (C4) to 91 h (B3). However the mortality of B20 was very low. With this exception, we can be confident that none of the SSIs had reduced virulence compared to their parent cultures, which is important if any of these isolates were to be taken forward by industry for development as a commercial product. The control mortality was slightly higher than expected but this was probably as a result of the increased handling time, as the mites and bees were produced at Rothamsted Research but the bioassay was done at Warwick HRI.

3.1.2 Conidia biomass production.Work was done to develop a method for producing large quantities of fungal spores for use in experiments with bees and varroa. Four fungal isolates were selected for use: A1, B3, C4, and F10. Fungal spores used for biocontrol can be applied in liquids or as dry formulations. For this project, it was decided that a dry formulation was the most appropriate since (a) honey bees react adversely to being sprayed with a liquid, and (b) application of dry formulations can often be done without specialised equipment, which would be very suitable for most bee keepers, who are hobbyists.

Initially, the fungal isolates were cultured on SDA in Petri dishes. The intention was to harvest the spores from these cultures and mix them with a carrier compound to produce a powder for use in bee bioassays. Carriers are inert ingredients that have no inherent insecticidal properties, and are used to dilute the active ingredient. In addition they often have characteristics that influence efficacy, for example by ensuring even application or by enhancing shelf life by adsorbing harmful secondary metabolites excreted by the control agent propagules during storage (Moore & Cauldwell, 1997). Numerous inert and organic materials have been tested as stabilising agents for various fungal propagules, including cornstarch, mexican lime, rice flour, diatomaceous earth and clays (Couch and Ignoffo, 1981). Clays are particularly attractive as they tend to have a consistent quality and can be selected for characteristics that are favourable to the fungus, for example pH, micro-climate and cation exchange capacity (Ward, 1984). However, the addition of various clays did not improve the stability of Metarhizium anisopliae var. acridium conidia which had been pre-dried to 7% moisture content (Moore & Higgins, 1997). Preliminary work in this project indicated that the clay carriers diatomaceous earth and kaolin had no effect on the viability of conidia of isolates A1, C4, and F10 when mixed in at a rate of 10: 1 (carrier: conidia), but that bentonite clay could reduce conidia viability (data not shown).

However, it became apparent early in the research programme that culturing the fungi on agar produced insufficient quantities of conidia for experiments, and hence it was decided to switch to a mass production system. We focused on the production of aerial conidia, which are the propagules that cause infection under natural conditions (Feng et al, 1994), and are the main reproductive phase of the fungal species used in this study. Most species of entomopathogenic fungi have hydrophobic conidia which are relatively resistant to harmful environmental factors (Jaronski, 1997). A second type of inoculum, blastospores, can also be cultured during vegetative growth in liquid culture. Blastospores are thin walled hyphal bodies, and although they are

SID 5 (2/05) Page 7 of 33

infectious and can germinate faster than aerial conidia (Jenkins & Thomas, 1996), they are more sensitive to desiccation (Bidochka et al, 1987; Kleepsies & Zimmerman, 1994). The literature suggests that aerial conidia can be produced more efficiently than blastospores, which along with their better environmental stability makes them the first choice for most commercial products (Wraight et al, 2001).

A wide range of substrates are available for mass production of aerial conidia including nutritive (natural) substrates and non nutritive substrates (e.g. clay granules coated with nutrients). For this project, it was decided to produce conidia on rice, which is reported to be superior to other nutritive substrates (Mendonca, 1992). In particular, the high surface area to volume ratio of rice leads to better nutrient absorption, gas exchange and heat transfer (Bartlett & Jaronski, 1988). The main advantage of non nutritive substrates is that the addition of nutrients can be optimised to maximise spore production (Jenkins & Lomer, 1994). However, it was concluded that, for this scale of project, the low cost and ease of use of rice as a substrate outweighed any disadvantage in terms of lack of control of nutrient content. In preliminary work we found that other cereal substrates, such as flaked maize or cracked wheat, have a tendency to clump after autoclaving and this can significantly impair fungal growth and sporulation. However, we found that clumping could be prevented if we used American long grain rice that was parboiled and then rinsed thoroughly prior to autoclaving, in order to wash away starch on the outside of the grains. Addition of calcium carbonate and calcium sulphate also helped reduce clumping.

A two stage system was used in which fungal mycelium was first produced in liquid culture using shake flasks, and then transferred to the solid substrate for conidia production. This system combines the benefits of high biomass production in the liquid culture stage with production of the stable aerial conidia on the solid substrate. It has a number of advantages, in that colonisation of the substrate is faster, and the competitiveness of the fungus is enhanced. The prepared American long grain rice rice substrate was contained within autoclavable bags used for the commercial production of mushroom spawn. These bags have a semi permeable membrane that enables gaseous exchage and thus improves aeration of fungal biomass.

For the rice grain system, conidia can be harvested in water (with a suitable wetting agent), but this may initiate germination or cause leaching of key compounds, and the subsequent removal of the water is reported to have a major negative impact on the viability of conidia unless done with extreme care (Burges, 1998). Therefore, it was decided to harvest the conidia in a dry system. Conidia production was terminated by drying the rice substrate, which is necessary to allow harvesting and also prevents the growth of contaminants (Burges, 1998). Previous research has shown that, in most cases, pre-harvest drying at temperatures below 30oC is necessary to maintain spore viability (Moore et al, 1996). Drying at these temperatures takes time, but the literature indicates strongly that it is essential to ensure inoculum quality. Hong et al, (2000) found that the optimum pre-harvest drying time for conidia of Metarhizium anisopliae mass produced on rice was 5 days in a forced air cabinet at 25 2oC and 45-75% RH, which reduced the moisture content of the substrate to 27-32% prior to harvest. We adopted a similar system, drying the conidia and substrate in a fume cupboard at room temperature.

The conidia of species C could be separated easily from the rice grains by sieving. However, the conidia of the other fungal species examined were situated within the main mass of mycelium on the rice grains and we found that it was not possible to obtain them by sieving. Instead, the dried fungal inoculum and rice grains were ground in a blender to produce a combination powder in which the rice grains acted as a carrier (Couch and Ignoffo, 1981). This obviated the requirement to investigate in detail the use of clay based carriers in this programme (see above), although they could have a role in a future commercial product. A machine for harvesting conidia based on cyclone extraction technology (called a Mycoharvester) has been developed by CABI Bioscience UK, which might enable conidia of species A, B, D, E or F to be separate from rice grains (either directly or following grinding in a blender). The equipment is expensive and was not available to this project, although it might well be appropriate for the production of a commercial control agent.

The finalised method developed in this project for conidia mass production was as follows: Batches of American long grain rice (1 kg) were washed twice in tap water to remove starch, parboiled

in 1 litre deionised water for 2 - 3 minutes, rinsed twice in tap water, then hand mixed with 50g CaCO3

and 50g CaSO4. Aliquots of 150 g were then added to 5 l mushroom spawn bags (Van Leer Packaging Systems Ltd, Poole, UK) that had been cut in half. The bags were sealed with masking tape and autoclaved at 121C for 30 minutes.

Fungal isolates were grown as SDA Petri dish cultures as described in 3.1.1. SDA liquid SDA media (100ml in 250ml conical flask) were inoculated with 1 x 108 conidia per flask, and maintained in an orbital shaking incubator in darkness at 23ºC and 200rpm for 3 days. (NB. lower shaker speeds resulted in the ‘balling’ of cultures). Biomass yield per flask (dry weight) was estimated from 50 ml sub-samples. Fungal material was collected by centrifugation (4660rpm for 10 minutes), washed in deionised water, and dried for 24 h at 100C.

Bags of prepared white rice were inoculated with 15ml of liquid culture, resealed, shaken thoroughly and incubated at 23C (species A, C, D, E, F) or 27C (species B). The bags were shaken by hand

SID 5 (2/05) Page 8 of 33

every 3 days. Production was terminated at 14 days by opening the bags in a fume cupboard (22C, 31% RH). Grain

samples (10 g) were then collected. The water content of the samples was calculated by drying for 24 h at 100C. Conidia yields were estimated from 5g grain samples which were mixed with 10 ml 0.03% Triton X-100, shaken vigorously, and enumerated using an improved Neubauer haemacytometer. Conidia viability was estimated by plating 20µl of a conidial suspension (1x 106 ml-1) onto SDA and recording the numbers of ungerminated and germinated conidia following incubation for 24 h at 23C (A, C, F) or 27C (B).

The grain in the opened bags was air dried for 5 days. Bulk processing was then done as follows: for C4, conidia were collected from grain by sieving (hand sieve, 200mm diameter, 160 µm pore size). For the other fungal isolates, material was ground for 1 minute at high speed in a Warring blender. Water content and conidia viability were then measured as described previously.

The system was evaluated using a total of six independent production batches, i.e. each batch was produced on a separate occasion.

Results of the rice based production system are summarised in Table 3. Average conidia yields ranged from 4.01 x 107 g-1 substrate (A1) to 5.88 x 108 g-1 substrate (C4). These yields are acceptable for experimental purposes, but higher yields would be desirable for commercial production. It may be possible to increase yields by further manipulation of the production process, for example by precise control of the water activity content of the substrate (Dorta et al, 1990), or by use of a nutritive substrate (see above). There are also options to search for high yielding isolates, which can return up to 1010 conidia g-1 of rice in mass production (Mendonca, 1992). The viability of conidia of all fungal isolates was very good at all stages of the production process (Table 3). The average moisture content of the grain after 14 days incubation was 19.8 % (s.d. 6.28%), and after 5 days air drying it had reduced to 3.9 % (s.d. 2.90 %) (Table 3). This is lower than that reported by Hong et al, (2000) for air drying of M. anisopliae produced on rice. It was an encouraging result, because low humidity is known to be important to the stability of conidia in storage (see below). It also meant that an additional post harvest drying step (which is used, for example, in the mass production of conidia of the M. anisopliae product ‘Green Muscle’, used for control of locusts and grasshoppers). Drying reduces the metabolic activity of the conidia which slows down the loss of storage reserves and minimises the production of toxic metabolites. The conidia moisture content of M. anisopliae, for example, needs to be reduced to around 5% for optimal storage (Moore et al, 1996).

3.1.3. Shelf life of mass produced conidiaA fungal control agent of varroa needs to have a good shelf life in order to be commercially viable. We believe that a fungal product is likely to be most useful when applied in the late summer / early autumn, after honey has been collected, in order to keep the bee colony in a good enough condition and size to survive winter (the period when bee colonies are most likely to die out as a result of varroa infestation). However, bee keepers might want to use the product at any time in spring, summer or autumn (i.e. when bees are foraging; they will not apply a product in the winter as opening up the hive could seriously damage the colony). This means that a bee keeper might end up using a product many months after it has been purchased. For entomopathogenic fungi used as microbial control agents of plant pests, it is generally accepted that a formulation should be stable for 12-18 months without refrigeration to meet market requirements (Couch & Ignoffo, 1981). As indicated above, low relative humidity is an important factor in the moderate-temperate storage stability of several entomopathogenic fungi. Diluents or carriers can be important also for protecting the fungus during storage against harmful products of metabolism and adverse environmental factors (Moore & Cauldwell, 1997). For example, the M. anisopliae product ‘Green Muscle’ is stored in mineral oil at a conidial moisture content of 5 % (Moore et al, 1996; Moore & Higgins, 1997). The mass production system used in the present study yielded conidia powders with an average moisture content of 3.9%, which was expected to give conidia with good longevity.

In this experiment, the viability of conidia produced using the rice-based system (see 3.1.2) was measured following storage for up to 48 weeks, in order to provide baseline data on the storage characteristics of candidate isolates. Three of the same fungal isolates were used as for Experiment 3.1.2: A1, C4, and F10. The fungi were grown on autoclaved rice in mushroom spawn bags for 14 days as described previously, after which the bags were opened and air dried for five days. The contents were then ground for 1 minute at high speed in a Warring blender (note that, in 3.1.2, conidia of C4 were extracted from the rice substrate by sieving. However, in this experiment, all isolates were processed in the same way for comparability). Batches of powder (2 g) were then placed in aluminium foil -coated seed bags (80 x 110 mm, McFarlane Group UK) and sealed with masking tape. The bags were stored for 0, 7, 14, 28, 56, 84, 112, 140, 168 and 336d at 4ºC, 22ºC (in a cooled incubator) or -20ºC. At each time interval, two replicate bags were removed and 1 g of conidia powder added to 10 ml 0.05% Triton X-100, vortex mixed for 30s, then diluted 1:10 (or 1:100 for C4). Aliquots (100 μl) were applied to three marked areas of an SDA plate, which was incubated at 23ºC in darkness for 24 h, after which the numbers of germinated and ungerminated conidia were counted (> 300 conidia per plate). The proportion of the conidia population that germinated was used to indicate the conidia viability. The experiment was done with three independent batches of conidia, i.e. produced on rice at different times. The data for the

SID 5 (2/05) Page 9 of 33

viability of conidia against length of storage period were analysed using a generalised linear model assuming a binomial distribution with a complementary log-log link function. There was evidence of over-dispersion in the data which was accounted for. The storage times required for 50% and 5% of conidia to remain viable (i.e. the viable storage times in days, VT50 & VT5) were calculated for each fungal isolate / temperature combination, together with 95% fiducial limits (Table 4).

The storage of conidia of all three isolates was improved as the storage temperature was reduced. However, none of the isolates showed evidence of being stable at room temperature for the 12 – 18 month target period that would be required for a commercial product (Couch & Ignoffo, 1981). Isolates A1 and F10 behaved similarly, with VT50 values of 202 d and 140 d respectively at 20ºC. Isolate C4 survived poorly, with a VT50 of only 7 d at 20ºC. It is likely that the shelf life of conidia of these fungal isolates could be improved considerably and these results show that further work is required in this area in order to develop a viable commercial biocontrol agent. The most likely causes of poor viability during storage are the production of toxic fungal metabolites and depletion of conidial reserves. It is possible that the moisture content of the conidia powder should be reduced even further by an additional post harvest drying step (this is in contrast to the indications from Moore et al. (1996), based on M. anisopliae, that the conidia moisture content for optimal storage needs to be c. 5%). Storage under a controlled atmosphere and addition of an inert carrier could also help. Fortunately, the finding that conidia viability can be modelled using the same distribution and link functions for different temperatures means that future storage studies can be fast tracked by running experiments for shorter periods and at elevated temperatures (although it will be necessary to avoid thermal stress. In this respect, the detailed models of fungal thermal biology developed in the previous project HH0813SHB could be used to identify a suitable study temperature). The finding that different fungal species / isolates varied significantly in the rate at which conidia lost viability during storage is also important for the future development of a commercial control agent. Isolate C4 appears to have a poor storage life compared to isolates of the same species used in commercial products, which suggests that variation within a species is likely to be as important as variation between fungal species. Hence, screening isolates for their storage characteristics is likely to be an important part of any future work to develop a commercial control agent.

3.1.4. Survival of conidia under bee colony conditionsSelf-evidently, a fungal control agent of V. destructor will have to function under the physical conditions of a honey bee colony (Chandler et al. 2001). Entomopathogenic fungi require conditions of high relative humidity (> 93% RH, see Andersen et al., 2006) and temperature conditions of 23 - 27ºC to infect their hosts, although some fungal isolates with higher temperature optima exist (see Davidson et al., 2003 for discussion). In contrast, the conditions within bee colonies are dry, averaging about 40% RH and rising to 70% RH during periods of evaporative cooling (Davidson et al., 2003). Fungal conidia exposed to these humidity conditions do not germinate. Bee hives are also warmer than the optimum temperature for many fungi, with temperatures ranging from c. 25ºC in broodless areas of the colony to 30 - 32ºC in the brood nest. In previous research (HH0813SHB) and in the current project, we have shown that entomopathogenic fungi can infect varroa mites under the environmental conditions found within honey bee colonies. In terms of relative humidity, it is likely that the mite’s body has a microclimate with a sufficiently high humidity for conidia germination even though the ambient humidity of the surrounding environment is low (varroa mites are hairy, and have a distinctive carapace which may provide high humidity conditions on the ventral surface). However, the dry environmental conditions within the bee colony could have a negative effect on those conidia that do not directly contact a varroa mite’s body following application to the hive. This could reduce the efficacy of conidia acquired by secondary pick up, which we have shown to be an important route of infection for other mite species (Chandler et al., 2005).

In Defra project HH0813SHB, we developed a method for measuring the survival of conidia under different conditions of temperature and humidity, using fungal isolate C4 as a model (this isolate was the most virulent to varroa in laboratory bioassays done at that time). Survival was measured at four combinations of temperature and humidity likely to be encountered in bee colonies: (a) 25C / 40 % RH; (b) 25C / 70 % RH; (c) 32.5C / 40 % RH; and (d) 32.5C / 70% RH. Conidia were maintained under these conditions for different lengths of time (during which they were unable to germinate because of the low humidity), and then transferred to high humidity conditions for 24 h, during which time the germination of the conidia was measured. The upper temperature used in the experiment represents that of the drone brood rearing area in a bee colony in summer (and matches the temperature optimum for V. destructor development) while the lower temperature represents that in broodless areas of the colony in summer. The relative humidities used in the experiment represent the average in brood areas in summer (40 % RH) and the maximum in brood areas (70% RH) during bouts of evaporative cooling. The results showed that the survival of conidia was reduced at low humidity (40 % RH compared with 70 % RH) and high temperature (32.5 compared to 25C). Even under the most favourable temperature / humidity combination (25C / 70 % RH), conidia did not survive longer than an estimated 380 h. These humidities are not sufficiently high to enable germination

Experiment 1: comparison of survival of 10 fungal isolates

SID 5 (2/05) Page 10 of 33

The survival of conidia of 10 fungal isolates (Table 1) was measured following exposure to one temperature / humidity combination (32.5C, 40% RH, which represents the harshest regime observed in the experiment done in HH0813SHB described in the previous paragraph) for 39.5h (the average survival time for conidia of C4 at these conditions estimated previously). Fungal isolates were grown on SDA as described in 3.1.1. For each isolate, a loop of fungal material was taken from an SDA culture and streaked onto a polycarbonate membrane (Anopore, 25mm diameter) and placed in the lid of a Petri dish (50mm diameter), with three replicate membranes per dish. Petri dish lids (and membranes) were placed on a stainless steel rack within a sandwich box which contained a saturated solution of potassium carbonate (400g per 250 ml distilled water) giving a relative humidity within the box of 40 % RH, and maintained at 32.5ºC. Membranes were removed after 39.5h and transferred to the surface of SDA in Petri dishes and maintained in darkness at 23OC for 24 h. Membranes were then placed on microscope slides and fixed with lactophenol methylene blue. Numbers of germinated and ungerminated conidia were recorded for approximately 100 conidia per membrane. Controls consisted of conidia that were placed on polycarbonate membranes straight onto SDA at 23ºC for 24 h. The experiment was repeated three times. For each isolate, the average percentage germination was calculated as a proportion of the control. The percentage germination of most of the isolates was similar (Table 5), and was also in agreement to the germination response of species C modelled previously in HH0813SHB. However, isolates A1 and A2 had a higher than expected survival, inferring a biologically-significant difference in conidial physiology.

Experiment 2: survival of species A isolatesA second experiment was done, in which the survival of isolates A1 and A2 was examined in greater detail, alongside C4 as a standard. Conidia germination was measured as described previously but this time conidia were maintained at 32.5ºC / 40% RH for periods of 0, 6, 12, 24, 36, 48, 60, 72, 96, 120, 144 and 168h. The experiment was repeated three times and the results combined and analysed using a generalised linear model incorporating a complementary log-log function (Genstat 2000). IGT50 values were then estimated, representing the incubation period under non-permissive conditions (i.e. 32.5ºC / 40% RH on polycarbonate membrane with no nutrients for germination) causing 50% germination of conidia when transferred to permissive conditions (i.e. 23ºC / 100% RH on SDA). The survival of conidia of A1 and A2 were similar to each other and were markedly higher than that of C4 442.99 (Table 6). The IGT50 values increased from 45 h (C4) to 164 and 183 h (A1 and A2). The IGT50 value obtained in this experiment for C4 was longer than that found in project (HH0813SHB) (22.4 (95% C.L. 19.84 –25.49)), however the pattern of response was similar (data not shown) and the difference was probably a result of small differences in the temperature and humidity during the experiment. Under these conditions (i.e. those likely to be found in the brood area during summer) conidia of C4 and A1 and A2 would not survive longer than an estimated 270, 501 and 435 h respectively. Hence species A is likely to persist within the hive for longer that species C or F, which is an important trait to consider when selecting fungal isolates for development as a commercial control agent

Summary Single spore isolates: With the exception of isolate B20, we can be confident that none of the SSIs had

reduced virulence compared to their parent cultures. Mass production: The yields of conidia in the mass production system were acceptable for experimental

purposes, but higher yields would be desirable for commercial production. Species C is probably the best for mass production, because conidia can be extracted easily from the substrate. The viability of conidia of all fungal isolates was very good at all stages of the production process.

Shelf life of mass produced conidia: None of the isolates showed evidence of being stable at room temperature for the 12 – 18 month target period that would be required for a commercial product. It is likely that the shelf life of conidia could be improved considerably by advances in processing and storage and this will be an important area for future work.

Conidia survival under simulated bee hive conditions. Laboratory experiments indicated that conidia of species A are likely to persist within a bee hive for considerably longer than those of other fungal species, which is an important trait to consider when selecting fungal isolates for development as a commercial control agent

3.2. (Objective 2): Effect of fungus in varroa and bee populations. Quantify the activity against varroa mites of fungi applied to bees.

3.2.1 General Methods:Combs containing mature worker bee pupae were removed from varroa-free colonies and held overnight in darkness at 35˚C. Newly emerged adult bees (<24h old) were transferred to wooden cages (25 x 25 x 110mm) and supplied ad libitum with water, pollen and 65% sucrose solution in gravity feeders (Bailey, 1971). Bees were held in groups of 15-30 as specified and incubated at 30˚C and ambient humidity. Adult female V. destructor were collected by hand from sealed worker brood from infested honey bee colonies. White-eyed bee pupae were collected by hand from uninfested colonies as a food source for the mites. One pupa and five mites were placed in each 1.5ml Eppendorf tube plugged with dental roll. Tubes were placed horizontally in

SID 5 (2/05) Page 11 of 33

the incubator in darkness at 30˚C for up to two weeks prior to experiments (Shaw et al, 2002). Fresh pupae were supplied every 5 days.

3.2.2. Effect of fungi on bees (feeding experiments).Feeding conidia suspensions to newly emerged adult beesIn previous research (HH0813SHB; see Shaw et al., 2002) fungal isolates were evaluated for their pathogenicity to A. mellifera in laboratory bioassays in which fungal conidia were applied topically to bees (this is the normal route by which entomopathogenic fungi invade their hosts). The results of this bioassay enabled isolates to be identified that did not infect bees by conventional contact action. However, it is also possible that bees could ingest fungal conidia applied to a bee hive. In this case, a fungus could have a deleterious effect on the bee either by the action of fungal metabolites in the gut (entomopathogenic fungi rarely invade the host by germination and outgrowth through the gut, so this mechanism of infection is unlikely) or by germination on and penetration through the mouthparts. There is a lot of oral contact between bees in a colony, for example when adult nurse bees feed and tend larvae, and when adult bees transfer food from one to another. Hence the possibility that fungal conidia could have a negative effect on bees as a result of ingestion needs to be considered. The aim of this section of the research project was to investigate the effects of fungi fed to bees.

The experiments done in this section of the project followed a worse case scenario, in which bees were fed a suspension of fungal conidia in a no-choice bioassay. Conidia suspensions of fungal isolates were prepared to required concentrations in 0.03% Tween 80. Caged bees were starved of sucrose and pollen for 1 hr prior to the experiment. Individual bees were fed 2 x 1µl drops of conidia suspension then returned to cages and incubated. Control bees were fed 2 x 1µl 0.03% Tween 80. A preliminary test showed that 0.03% Tween had no effect on the survival of adult bees at this dose. Dead bees were removed, counted daily and total mortality determined. Although cadavers were incubated at high humidities to observe outgrowth of fungus, indicating fungus-induced mortality, the levels of sporulation were variable, and hence total mortality only was used in analysis. This approach was taken in all experiments considering bee mortality.

A preliminary experiment was done in which the susceptibility of bees to isolates of four species: A1, B3, C4, and F10 was assessed at a concentration of 6 x 108 conidia ml-1. Four replicate cages of 25 bees were fed per treatment. Mortality was assessed for 16 days post inoculation. Isolate C4 reduced the survival of bees more compared with both the control and the other isolates. Bee mortality was 100% within 12 days with this isolate. Levels of bee mortality at 16 days for other treatments were approximately 5% (control); 10% (B3), 10% (F10) and 20% (A1).

The preliminary experiment was followed by a large scale study, in which the susceptibility of bees to conidia of A1, A2, C4, C5, C6, C7, E9, and F10 was assessed at a concentration of 1 x 108 conidia ml-1. The experiment was done according to a randomised block design with two blocks of nine treatments (eight isolates plus a control) run on each of three occasions. Hence, a total of six cages of 25 bees were fed per treatment. Additionally, to ensure that any differences in efficiency of the three people feeding the bees were not confounded with treatment differences, three cages in each block were allocated to each person according to a 3 x 3 Latin square so that over the first three blocks (and again over the second three) each person fed bees for one replicate of each treatment. Total mortality was assessed for 15 days post inoculation as described previously. No bees died in the control cages. Weibull distributions with three parameters (shape, scale, and lag where appropriate) were fitted to the cumulative mortality counts for five of the eight isolates and LT50s (lethal time to kill 50% of bees) determined. The findings of the experiment have been grouped according to fungal taxa as follows: Species D – F isolates. LT50s were not quantified for E9 or F10 because bee mortality never exceeded

10-30% during the experiment. Therefore these isolates were considered to be non pathogenic to adult bees when fed as suspensions.

Isolates of species A & C. Data for isolate A1 was excluded because four of the six replicates failed when the plates did not sporulate sufficiently. This left insufficient data for further analysis. Bee mortality for isolate A1 exceeded 90% in two replicates. Bee mortality in the four species C (C4 – C7) reached 100% in at least one replicate and varied between 40-100%. All these isolates caused. Mean LT50s for the five isolates (C4 – C7 and A1) were compared using regression methods. Isolate C6 had a significantly later LT50 of 8.8 days (P < 0.001) compared to the remaining four isolates, which were not significantly different to each other (P = 0.497) and had a combined mean LT50 of 7.0 days. Final mortality at day 15 as a proportion of the initial number of bees per cage was compared between these five isolates (all replicates) using a Generalized Linear Model (GLM) with binomial error and logit link and allowance for overdispersion. No difference was observed amongst the five isolates (P = 0.632), the overall proportion mortality on the logit scale being 1.72 (SEM = 0.286, n=30; backtransformed value = 82%). For the isolates that were run in both the preliminary and the full experiment, the results were very similar.

It should be emphasised that this route of exposure of bees to fungal conidia is unlikely to occur in a bee hive. Nevertheless, it indicates that isolates of species A and C can cause mortality per os in bees, although the

SID 5 (2/05) Page 12 of 33

mechanism of action is unknown. In contrast, species D – F isolates were non pathogenic. In previous work, these isolates caused no mortality when applied topically to bees in laboratory bioassays.

Feeding fungi in pollen to newly emerged adult bees A more likely way in which adult bees could ingest fungal conidia is with pollen. Bees treated with a powder of conidia are likely to engage in grooming activity and in so doing transfer conidia to pollen baskets on the hind legs. This pollen is then transferred to the bees to pollen stores in the hive. Conidia may then be ingested when bees feed from the pollen stores. In this experiment, bees were caged, 25 per cage and incubated at 30˚C and ambient humidity. Preliminary investigations showed that a cage of 25 bees would consume, on average, 0.15g pollen per day. In the following experiments conidia were mixed with 0.15g pollen per cage to ensure that the dose was consumed within 24h. Conidia and pollen were combined and packed tightly into pollen feeders.

A preliminary experiment was done using conidia powder produced on rice based substrate (3.1.2). 1 x 109 conidia of A1, A2, C4 – C7, E9 and F10 were mixed with 0.15g pollen per cage of 25 bees. A large bulk of conidia powder was needed to provide a dose of 1 x 109 conidia. This, when combined with pollen, produced a powder more loose than pollen. These pollen mixtures were repulsive to bees, with the exception of F10, and so the methodology was adapted for the subsequent large scale experiment.

The large scale experiment was done as follows: Conidia suspensions of fungal isolates (A1, A2, C4 – C7, E9 and F10) were prepared in 0.03% Tween 80 from cultures grown on SDA. 1ml aliquots were prepared at 1 x 109, 1 x 108 and 1 x 107 conidia ml-1 in 1.5ml eppendorf tubes and centrifuged for 5 min at 12000rpm. Tween 80 was decanted off and 0.15g pollen added and mixed with the spore pellet that remained in each tube. The lid of the eppendorf was removed and the tube introduced to the cage of bees in replacement of the standard pollen feeder. Two cages of 25 bees were treated per isolate per dose plus two control cages. Cages were placed into separate trays and incubated. Bees were provided with untreated pollen for two days prior to the experiment and ad libitum post inoculation. Control bees were fed 0.15g untreated pollen per cage. Control mortality ranged between 2.1 and 9.3%. Dead bees were removed and counted daily. Isolate C6 was still completely repulsive to bees at a dose of 1 x 109 conidia using this method. Consequently all bees in both replicates died rapidly of starvation and no data on susceptibility were available. At the same dose (1 x 109 conidia) isolates C5 & C7 were sufficiently repulsive that bees only consumed contaminated pollen in one of each of the two replicates within the first three days and so the results represent only a single replicate for each isolate. Indeed, at the highest dose most of the species C isolates were repulsive to bees and so, in general, they avoided consuming them. However, on the occasions when they were consumed (isolates C5 and C7), significant mortality of bees (i.e. > c. 10%) was only caused by isolate C5 and even then 30% of bees survived. For all other species and over all three doses evaluated mortality of bees was never greater than 20%. This confirms the previous finding that isolates of species D – F were non pathogenic to bees. The reason for the observation that isolates of A & C caused significant bee mortality when fed to bees in liquid suspension, but caused no mortality when fed in pollen, is unknown. It is possible that the liquid suspension provided free water sufficient for conidia to germinate and infect mouthparts. It is also possible that the pollen had antimicrobial properties. The fact that some of the species C isolates were repulsive to bees when fed in pollen suggest that bees are unlikely to expose themselves to inoculum by feeding if these isolates were used on a hive scale. The concern would be that if conidia of these isolates were present in pollen stores in large quantities, then they could have a negative effect on the colony by inhibiting pollen feeding. This is an area that requires further study but would be an important factor to consider when it comes to fungal isolate selection.

Feeding conidia suspensions to honey bee larvaeHoney bee larvae may differ in their susceptibility to entomopathogenic fungi compared to adult bees. The routes of exposure to fungal inoculum are likely to be different for larval bees. Larvae are attended by adult nurse bees and it is possible that they could be exposed to fungal conidia acquired by nurse bees when the hive is treated. In this experiment, honey bee colonies were confined in "pollination cages" to prevent drifting of bees between colonies. Cages were placed on grass approx one metre apart, one colony in each cage. Cages consisted of frames of galvanised steel (3.66m x 8.23m x 2.74m) covered in netting (Clear Saran fabric T0984/000, 0.78 x 0.38 mm mesh, Simpers Ropeworks Ltd, Cambridge, UK). 65% sucrose solution and fresh pollen were supplied ad libitum on a feeding table located at the corner of each cage, opposite the bee colony. The sucrose was supplied in a von Frisch feeder and the pollen, which had been collected from foragers and frozen for storage, was defrosted and presented in a glass Petri dish. Colonies were allowed to establish for 1 week prior to the experiment. Brood combs were removed one at a time for feeding. Approximately 50 larvae (in 5 batches of 10 contiguous cells) were fed on each comb for each treatment applied, and combs had either two or three treatments allocated to them: (1) two controls (control colony), (2) two controls and species C (treated colony), (3) two controls and species F (treated colony). For fungal feeding, individual larvae, 2-3 days old, were fed 2 x 1µl drops of conidia suspension administered to the brood food at the bottom of the cell. Control larvae were fed either 2 x 1µl 0.03% Tween 80 (control 1), or nothing (control 2). A preliminary test showed that 0.03% Tween had no effect on the survival of larvae or emergence of adult bees at this dose.

SID 5 (2/05) Page 13 of 33

Brood combs were returned to the hive immediately after feeding. In these experiments it was not possible to collect removed larvae and so we could not determine whether larval mortality (which would have resulted in removal of the dead larva by the nurse bees) was due to fungal activity or other factors. The number of larvae removed indicated the overall level of mortality while the number of capped cells represented those larvae surviving to the pupal stage.

In an initial experiment two treated colonies were set up for each isolate, plus two control colonies. Test larvae were fed suspensions of C7 and F10 at 1.5 x 108 conidia ml-1. Larval cells were checked five days post inoculation and removal of larvae / capping of cells was recorded. The experiment was repeated on a second occasion, however colonies did not establish well in the pollination cages, so only 1 treated colony per isolate was possible, plus one control colony. Test larvae were fed suspensions of C7 and F10 at 1 x 108 conidia ml-1 as described above. Larval cells were checked at day five for larvae removal / cell capping as before, and again at emergence (approx day 21) to determine whether the developing bee had survived after capping. The numbers of capped cells after five days (as a proportion of the number of larvae treated) were combined from the experiments to give three replicates of each of eight treatments in total : (1) mortality in cells treated with fungal species C; (2) mortality in cells treated with fungal species F; (3) mortality in control 1 cells; (4) mortality in control 2 (Tween-treated) cells; (5) mortality in control 1 cells adjacent to cells treated with species C; (6) mortality in control 1 cells adjacent to cells treated with species F; (7) mortality in control 2 cells adjacent to cells treated with species C; (8) mortality in control 2 cells adjacent to cells treated with species F. The mean proportions of capped cells were compared amongst treatments using a Generalized Linear Model with binomial error and logit link and allowance for overdispersion as described earlier. The proportion of larvae capped was similar for the F10 treatment, the control larvae from the same comb and the true controls in the absence of the fungus entirely (Figure 1). However, the proportion of larvae capped in the C7 treatment was significantly smaller, both for larvae fed conidia directly in the brood food and larvae in the adjacent control cells on the same comb. This indicated that C7 fed larvae were less likely to survive and be capped than those fed F10 and that the fungus was likely to be transferred to adjacent control larvae by nurse bees, causing similar mortality. Overall emergence of capped bees at day 21 (second experiment) was 87% (s.e. 2.0%) and was similar across all treatments suggesting that larvae that survived long enough to pupate and be capped had escaped infection. It is possible that C7 was repulsive to larval bees, in the same way that this isolate deterred feeding in adult bees (see above), which if this occurred would have caused larval mortality. Infection of larvae is also possible, although the finding later in the project that species C conidia survived poorly in honey (see 3.3.2) suggests that further work needs to be done to examine fungal survival in brood food. The finding that entomopathogenic fungi are spread by nurse bees to larvae could provide a mechanism for spreading the fungus to varroa mites on brood, although the isolates used would have to be ‘bee friendly’.

Feeding carrier powders to newly emerged adult bees As described previously, many commercial fungal biocontrol agents are formulated in inert carriers. Therefore it is necessary to investigate such carriers for any deleterious effects on bees. In this experiment, bees were caged and provided with untreated pollen for two days prior to the experiment and ad libitum post inoculation. The four carrier powders tested were kaolin clay, bentonite, talc and diatomaceous earth. 1g pollen was combined with 0.25g carrier and packed tightly into the pollen feeder. Four cages of 25 bees were treated per carrier plus four control cages of 1g pollen. Cages were incubated and pollen consumption monitored. All pollen was consumed within the experimental period (24 days). Dead bees were removed and counted daily. There were no observable deleterious effects on bees with any of the carrier powders tested as mortality was no more than 10% for any of these four treatments. In fact, in this experiment, control mortality was greater than for all carrier treatments combined (P =0.032), and there were no differences amongst the four carriers (P = 0.097) based on a Generalized Linear Model for number dead after 24 days as a proportion of the number bees tested, with binomial error and logit link (Table 7).

3.2.3 Dissemination of fungus from bees to mites.While a proportion of fungal inoculum applied to a bee hive will contact varroa mites directly (these will be mainly varroa mites feeding on adult bees although a small proportion of mites will be free moving in the hive at any one time), secondary acquisition of inoculum from the surface of adult honey bees is also likely to be an important (indeed maybe the most important) route of exposure. In this part of the research project, work was done to investigate the dissemination of fungal conidia from bees to varroa mites.

Mortality of mites on bees dusted with conidia Preliminary experiment to develop methodology and to determine whether bees contaminated with conidia of entomopathogenic fungi can indirectly initiate infection in mites placed on them: Cages of 15 newly emerged adult bees were incubated for 10 days prior to the experiment. Conidia of A1, A2, C4, C7, E9 and F10 produced on rice substrate (as described earlier) were used. For each cage, the bees were anaesthetised with CO2 for 60s and transferred to a plastic jar with a screw lid. 1g conidia powder was added to the jar which was gently rolled to coat the bees. The bees were removed and returned to the cage, which was placed in a box and covered with filter paper to prevent movement of conidia between cages. Bees were allowed to recover at room temperature then the boxes were placed in the incubator. The weight of the residual powder was

SID 5 (2/05) Page 14 of 33

measured. The average dose applied per cage of bees was estimated at 2.0 x 108 conidia (s.e.m. 0.29 x 108). Two cages of 15 bees were dusted per isolate plus 2 control cages (bees handled in the same way but without dusting). After 24h ten varroa mites were added to each cage. Dead bees and mites (no response to stimulus) were removed and counted daily. In all the treatments mite mortality reached 100% and all these mites sporulated indicating mortality due to infection. However, mite mortality was also high in the control treatments, varying between 30 and 60%, although these did not sporulate indicating they were not infected. In addition, bee mortality was also high in all treatments including controls (25 % mortality in controls) suggesting that the experimental design required improvement. However, even given these problems bee mortality was greater and occurred faster in species A and C treatments than in the control and species E & F treatments which were similar to each other, suggesting that the E & F species were less aggressive to bees. High bee mortality may have been caused by the method of inoculation or the powder formulation. These aspects were addressed in the next experiment and hence no formal analysis was done on the data from this experiment.

Experiments to determine whether bees contaminated with F10 technical powder can indirectly initiate infection in mites placed on them: Three experiments were done. Newly emerged adult bees were caged in groups of 30 and incubated for 3 days prior to each experiment. Conidia were applied to bees in test cages. Each test cage was made from two 12.5cm square of 6mm Perspex, one with a 12.5 cm square of 0.711mm galvanised wire mesh with an aperture of 2.46mm fixed over a 10cm diameter hold. 0.025g F10 technical powder was placed on a 10.5cm diameter filter paper on the base of the test cage. Bees were anaesthetised for 60 s with CO2 and transferred onto the filter paper in the test cage. The two sides of the cage were held together with bulldog clips. The depth of the Perspex allowed the bees to move freely under the wire screen without clustering. Bees were allowed to recover and acquire conidia. Full exposure time was 45 min after which bees were returned to wooden cages without anesthetisation. In all three experiments control groups of bees were handled in the same way but received no application of powder. In experiments 2 and 3 groups of bees were also treated with non-pathogenic F10 (technical powder heated to 121˚C for 15 mins). For F10 and non-pathogenic heat treated F10 treatments the weight of residual powder on the filter papers was measured. The average dose applied per cage of bees was 1.03 x 109 conidia for F10 and 0.81 x 109 conidia for heat treated F10 powder (no significant difference between the two, P > 0.05). Approximately 20, 25 and 25 varroa mites were added to each cage in experiments 1-3, respectively. However, in experiments 2 and 3 the bees were naturally infested with mites prior to inoculation with conidia and so there were often more than the planned number of mites per cage. Preliminary work had shown that removal of the mites increased handling mortality in the bees and so they were not removed. Cages were then placed individually in aluminium food trays 16cm x 7cm x 5cm, sealed with a cardboard lid and incubated. There were five cages per treatment in experiments 1 and 3, and six of F10 powder and five of heat-treated F10 powder and controls in experiment 2. Dead bees and mites were removed and counted daily for ten days in experiments 1 and 2, and for seven days in experiment 3. Mite cadavers were placed on 1% tap water agar and incubated at 23˚C. The growth of sporulating mycelia on these cadavers was taken to indicate fungus-induced mortality. At the end of the experiment each unit was placed in the freezer (-20°C) overnight to kill any remaining bees and mites and determine the total mite population. Mite mortality was generally higher in experiment 1 compared to experiments 2 & 3 (Table 8). In experiment 1 mite mortality per control cage ranged between 20 and 50%. In experiments 2 & 3 mortality per cage over both the control and heat treated F10 treatments was generally less than 20% (i.e. for 15 out of 20 cages in total). In all the control and heat treated F10 treatments, none of the cadavers sporulated indicating they were not infected. Counts of total mite mortality (and the proportion of those that were mycosed) and of total bee mortality at day 10 (day 7 in experiment 3) post inoculation were analysed using binomial GLMs as described earlier. There was significantly greater mortality in the F10 treated mites than mites treated with heat treated F10 or the controls in all three experiments (P < 0.01). Furthermore, there was no significant difference between mite mortality in the heat treated F10 and control treatment demonstrating that mortality in the F10 treatment was caused by the action of the living fungus, and not due to a physical effect of the powder. Of the cadavers dying in the F10 treatments between 49 and 97 % sporulated demonstrating that they died of infection and this was consistent across experiments (Table 9). Cumulative daily mite mortality for the F10 treatment counts were analysed as described previously to compute LT50s. The average unweighted LT50 over all experiments was 6.75 days (weighted = 5.13 days) and there was no significant effect of dose on LT50 (data not shown). No bee mortality occurred in experiment 1 and it never exceeded 27% in any of the other experiments demonstrating that bees did not suffer from being treated with conidia formulations and were not infected by F10 (Table 10)

Dissemination of fungi from foragers to nurse bees to mitesExperiment to determine whether forager bees contaminated with conidia of entomopathogenic fungi can transfer inoculum to nurse bees and from there indirectly initiate infection in mites: Newly emerged bees were collected and caged in groups of 15 and incubated for 10 days until they were old enough to have become forager bees. Conidia of A1, A2, C4, C7, E9 and F10 produced on rice substrate were used. Groups of the forager bees were anaesthetised with CO2 for 60 s and transferred to a plastic jar with screw lid. 1g conidia were added to the jar and the bees were inoculated as described previously. The dose applied ranged from 1.3 – 3.4 x 108 conidia per cage of bees. Two cages of 15 bees were dusted per isolate plus 2 control cages (bees handled in the same way but without dusting) and incubated for 24 h. More newly emerged bees (of an

SID 5 (2/05) Page 15 of 33

age that they would be nurse bees) were then collected and one varroa mite was placed onto each of these bees and approximately 15 such bees added to each cage of foragers. Dead bees (nurse and forager bees) and mites were removed and counted daily for 14 days. Mite numbers varied between 4 and 19 per cage suggesting that some mites were lost during the experiment or could have moved between treatments. Mite and bee mortality were analysed using a GLM as described previously. There were no differences amongst the total mite mortalities for isolates A1, E9, F10 and the control (group 1: P = 0.620, d14 mortality = 40% approx). Similarly, there were no differences amongst isolates A1, C4 & C7 (group 2: P = 0.245, d14 mortality = 90% approx.). However, there was a significant differences between these two groups overall (P <0.001), with group 2 causing more mortality in mites than group 1. Note that the amount of fungal inoculum applied was less than that used in the previous experiment (by a factor of 10). Bee mortality was high over all treatments and hence results should be treated with caution. There were no differences amongst the mortalities for A2, E9, F10 and the control (group 1: P =0.793, d 14 mortality = 50% approx.) and similarly no differences amongst isolates A1, C4 & C7 (group 2: P =0.737, d14 mortality = 80% approx.). However, as for mite mortality, there was a significant difference between these two groups overall (P =0.024). Group 2 caused greater mortality than group 1. This was a preliminary experiment and there was high bee mortality, perhaps partially due to the conidia being formulated in a rice substrate. A second experiment was attempted in which forager bees were marked to be able to distinguish them from nurse bees but on this occasion there was significant contamination with Aspergillus spp and the experiment was aborted.

3.2.4 Effect of fungal inoculum on varroa mite and honey bee populations within a small nucleus colony in a flight room fungus.A honey bee colony was established in a cage in a flight room. The cage consisted of a galvanised steel frame (3.1m x 3.1m x max. height 2.3 m) covered in netting (Clear Saran fabric T0984/000, 0.78 x 0.38 mm mesh, Simpers Ropeworks Ltd, Cambridge, UK) The room was maintained at a temperature of 25±1˚C, relative humidity of 60±5% and photoperiod LD 10:14h. 65% sucrose solution and fresh pollen were supplied ad libitum on a feeding table in the centre of the room. The sucrose was supplied in a von Frisch feeder and the pollen, which had been collected from foragers from field hives and frozen for storage, was defrosted and presented in a glass Petri dish. The colony consisted of: a central frame of empty cells with a square approx 5cm x 5cm cut from the centre; 2 x brood combs from colonies infested with V. destructor; 2 x food frames with sealed and unsealed honey plus pollen; approx 2000 honey bee adults and a queen from a healthy colony. The colony was allowed to establish for one week prior to experimentation. Prior to the first introduction of fungus, 2 x Viburnum sp., 1 x Chaenomeles cathayensis (Japanese quince) around the feeding table as sources of forage. In addition 2 x Brassica oleracea (ornamental cabbage) were placed in the flight room around the feeding table. Prior to the second introduction of fungus approximately two months later these were replaced with five oilseed rape (OSR) plants, Brassica napus cv. Heros, that were just beginning to flower. The first introduction of fungus was made on 21/11/05. Eight cages of 40 10 day old bees (total 320 bees) were inoculated with F10 technical powder using the method described earlier. Cages of inoculated bees were opened and placed on top of the flight room colony, under the lid to allow inoculated bees to move into the colony. Bees were inoculated with a total of 6.01 x 1010 conidia. The second introduction of fungus was on the 01/02/06. On this occasion 3 x 3.33g F10 technical powder was distributed on the top bars of the three brood combs in the colony. The total dose applied was 6.7 x 1011 conidia. Throughout the experiment the colony was fitted with a varroa screen (a solid wood surface that slides into the bottom of the hive to collect falling mites) and mite drop was counted daily prior and post inoculation. Post inoculation, dead mites were removed placed on 1% tap water agar and incubated at 23˚C. Outgrowth of sporulating mycelia on cadavers was taken to indicate fungus-induced mortality. Bees found dead on the floor of the flight room were counted and removed daily. Level of mycosis on dead mites was high immediately after both inoculations, indicating that mite drop during this period was due to fungal action, and decreased to zero after around 13 days on both occasions (Figure 2, 3). Although the level of mycosis was generally lower for the first inoculation, the pattern of change over time was similar for both inoculations. Approximately 50% of dead mites sporulated during the period of 18 – 19 days after inoculation (Table 11). There was considerable natural variation in the daily mite drop prior to the first treatment, which was also high relative to later in the experiment. This may have been a result of disturbance when the colony was established. It will be important in future experiments to account for this underlying variation in mite drop by running large scale experiments and by allowing colonies to establish for longer before treatment (the aim would be for colonies to achieve more of a steady state condition resulting in less variable natural mite drop, although the danger with a nucleus colony is that bees may rapidly fill up their allocated space, leading to stress in the colony). Nevertheless, the results indicate that F10 technical powder caused high levels of infection in the varroa population. Priorities for future research will be to refine the method used to assess fungal efficacy and to develop an efficient application system, including a method to prolong the persistence of fungal induced mortality in varroa. Kanga et al., (2003) obtained good control of varroa in a field trial based on 36 honey bee colonies, using M. anisopliae, which was applied as a dust or on plastic strips placed between the frames. However, when applied at a high concentration, the fungus caused a decline in the estimated population of bees. We have used the same isolate in our project and found it to be pathogenic to bees. At low concentrations, the fungus had less of an effect, although we would have some concerns about the long term impact of this isolate on bees.

SID 5 (2/05) Page 16 of 33

Summary Species E & F isolates cause no mortality to bees when applied topically or per os. Similarly, F10 had no

negative effect on larval bees. Although species A and C caused high levels of mortality in adult bees when applied per os in a liquid

suspension, they are unlikely to cause infection in the more realistic scenario of bees eating conidia-contaminated pollen. It is noteworthy that species C mixed with pollen was repellent to bees, and this could have a negative effect on bee feeding in a hive.

Bee larvae fed species C showed reduced survival, although larvae that survived to capping pupated as normal.

Varroa mites are able to acquire fungal inoculum from adult bees treated with F10 resulting in good levels of mortality.

F10 technical powder caused high levels of mycosis in dead mites recovered from a nucleus colony experiment. Future studies need to refine the methods used to evaluate fungal efficacy on the bee-colony scale, and to develop an efficient application system, including a method to prolong the persistence of fungal induced mortality in varroa. I

3.3. (Objective 3): Environmental impact : Quantify the dissemination of fungal conidia to bee products and the environment.

3.3.1 Detection systems (molecular markers / serology)Serology Polyclonal antisera were produced and methods investigated for the quantitative detection of fungal conidia in bee and mite populations and for the fluorescent labelling of conidia to identify fungal species active against varroa. Investigation of spore disruption techniques. Fungi generally produce more cross-reactive antisera than other antigens. Fungal mycelia are the most immunogenic and intact conidia are the least. As we were using conidia to inoculate varroa mites we needed to raise antisera to those conidia if we were to detect them on mites. Germinated or disrupted conidia have more antigenic sites and so produce a more specific antiserum. Germinated conidia of some isolates (species C particularly) may also produce toxins in mice. For this reason we chose to disrupt conidia for antigens rather than use germinated conidia. Disruption techniques investigated included sonication (blasting spores with sound waves); freezing and thawing; and FastPrep (homogenisation using glass beads). Small volumes of conidia suspensions were tested using these techniques and disruption checked visually with the microscope. The FastPrep machine (20s at speed of 6.5, 10min on ice, 20s at 6.5) produced the best results.

Production of polyclonal antisera to whole and disrupted conidia. In an attempt to maximise specificity, antisera were produced in mice rather than rabbits. Conidia suspensions of C4, A1, F10 and B3 were prepared in phosphate-buffered saline. Two 0.5ml aliquots of 1 x 109 conidia ml-1 were prepared for each isolate. One aliquot of each isolate was disrupted using the FastPrep machine as described in O3.11. The immunisation protocol involved two immunisations of 0.04ml of antigen in 0.2ml PBS 8 weeks apart, followed by three bleeds 4 weeks apart. Antisera were raised to whole and disrupted spores of each isolate.

Antigen-coated plate (ACP) ELISA to determine titre of antisera. ELISA procedures were investigated to determine antisera working dilutions and sensitivity of detection. Methodologies failed (no response from positive controls) possibly due to the enzyme system or small working volumes (100µl). Immunofluorescence was also investigated as an alternative method to ELISA, using lectins as a positive control. However, serological investigations were not pursued due to constraints of working with small volumes of antisera produced in mice and the time taken to produce large enough volumes for use in experiments. Molecular techniques for detection were, therefore, given preference.

Molecular markers DNA based techniques offer a method for the detection of fungi with high sensitivity. They can be particularly useful for the detection of fungi in environmental samples, where isolation of viable material using selective media can be problematic because of contamination by other microorganisms. In this study, nucleotide sequence information of the rRNA gene repeat unit (ITS I, 5.8S gene, ITS II) was obtained for isolates of Species A, B, C & F (10 isolates studied in total, all isolates were single spored, see Table 2) using PCR amplification with universal fungal primers (White et al.,1990) followed by sequencing. Forward and reverse primers were then designed for each fungus and evaluated for their ability to amplify DNA from the other fungi / isolates (these fungi occur in the same family and hence the possibility that primers from one species / genus could amplify DNA from another needed to be examined). The experiment was done as follows: Fungal isolates were grown in SDA liquid shake culture as described in 3.1.2. Mycelium was harvested by filtration through muslin, rinsed with sterile water, blotted dry then freeze dried and stored at -20ºC until required. DNA was extracted from 100mg freeze dried mycelium using a GenElute plant genomic DNA miniprep kit (Sigma-

SID 5 (2/05) Page 17 of 33

Aldrich, Poole, UK). The concentration of DNA was measured using a Nanodrop ND-1000 spectrophotometer (Nanodrop Technologies, Wilmington, USA). Fungal DNA (1ng) was amplified by PCR using ITS primers 1 and 4 (White et al.,1990). The thermocycler conditions were as follows: (a) Initial denaturing 94ºC for 2 min, annealing 55ºC 30s: (b) 35 cycles of extension 72ºC 30s denaturing 94ºC for 30 s, annealing 55ºC 30s; (c) final extension conditions of 72ºC 5 min. PCR products were then separated on a 1.5% agarose gel at 6V.cm-1 for 1h and visualised using ethidium bromide staining and exposure to UV light. PCR products were purified using a QIAquick PCR product purification kit (Qiagen, Crawley, UK) then a BigDye Terminator v3.1 cycle sequencing kit (Applied Biosystems, Warrington, UK) was used together with ITS primers 1 and 4 to generate forward and reverse products. Sequence data was produced by an ABI 3130xl genetic analyser (Applied Biosystems, Warrington UK). These sequences were compared and consensus versions were constructed. A multiple sequence alignment programme (MegAlign, DNASTAR Inc., Madison, USA) was used to compare these sequences and others downloaded from DNA databases available on the internet (National Center for Biotechnology Information, http://www.ncbi.nlm.nih.gov/). Putative genus-specific primers for species A - F were identified from these aligned sequences and checked for suitability in PCR with PrimerSelect (DNASTAR Inc., Madison, USA). The specificity of the primers was tested by using target DNA of all four genera in turn with each primer pair. PCR thermocycler conditions were as described above for ITS primers. Four pairs of primers were chosen as being specific (Table 12):

3.3.2. Impact of fungus on bee products: survival of fungal spores in honey Although our current thinking is that a fungal control agent of varroa would be used in the late summer / early autumn, after the honey crop has been collected, there is still a risk that fungal conidia could be transferred to honey, for example to uncapped honey stores left in place by the bee keeper over winter, or if the bee keeper uses the product before the honey harvest. In this part of the project, an experiment was done to measure the survival of fungal conidia in honey. It is difficult to predict the effects of honey on spore viability, because honey has antimicrobial properties, but is also known as a preservative, and hence spores deposited in it could remain viable for extended periods. At the same time, the effect of honey on the conidia of biocontrol fungi needs to set within context, because it is common for honey to naturally contain spores from a range of fungi transported by bees from the environment to the bee colony, which can include bee pathogenic fungi. Discussion of the potential for transport of conidia to honey cells is investigated in the next section.

The effect of honey on the viability of fungal conidia was investigated as follows. The experiment was done with three isolates: A1, C4, and F10. Conidia suspensions (1 x 106 conidia.ml-1) were prepared in 0.05% Triton X 100 as described previously from 7-14 day-old SDA cultures. 4μl of each suspension was added to a sterilised microfuge tube and then approximately 500μl of heat-treated (65C for 5 days) honey (Pure Country Honey, Rowse Honey, Wallingford, Oxon) was placed on top and tubes were vortexed to mix the fluids. Replicate tubes were then incubated at 5C and 20C. Three replicate tubes per isolate / temperature combination were removed at 0, 7 and 14 days. On each occasion, 500μl of 0.05% Triton X 100 was added to each tube and the contents mixed by vortexing. The contents of the tubes were then transferred to sterile 30ml universal bottles. Each microfuge tube was rinsed twice with 0.05% Triton X 100 and the washings added to the original contents in the universal bottles. The final volume was adjusted to 20ml with 0.05% Triton X 100 and 500μl aliquots were spread over SDA plates. Numbers of colony forming units per plate were recorded after 3-5 days at 23C. Survival was compared against controls using 0.05% Triton X 100 in place of honey with 0 days incubation.

Treatment with honey caused rapid death of conidia for A1 and C4 (Table 13). The data suggest that conidia of C4 died very soon after exposure to honey, while conidia of A1 did not survive longer than 7 days. The poor survival of C4 is in keeping with the poor survival of this isolate in long term storage (see 3.1.3). Conidia of F10 appeared to have more tolerance of honey, although 100% mortality was still observed by day 7 at 20C and by day 14 at 5C. It is possible that conidia were subject to rapid osmolytic stress following immersion in honey, which has a low water content and high water potential. On the other hand, the storage stability of conidia of commercial mycopesticides is known to be enhanced under low humidity conditions, which suggests that mortality could also have been caused by other mechanisms, for example the presence of antifungal compounds. Whatever the mechanism, the data suggests that fungal conidia applied to a bee hive are unlikely to survive if transported into honey.

3.3.3 Potential movement of fungus within a bee colony and in the environmentThe movement of F10 technical powder within a bee colony and onto flowers being visited by treated bees was investigated as part of 3.2.4. F10 powder was applied to bees from a hive within a flight room as described in section 3.24. The following samples were collected before and after application of F10 powder: (1) live foraging bees [LF]; (2) live brood nest adult worker bees [BN]; (3) pollen [P]; (4) honey [H]; (5) flowers of Brassica napus [F]. The before treatment samples were collected one day prior to application of F10 powder. For post treatment samples, collection of bees and flowers was done one day after application of F10, while honey and pollen samples were collected 21 days after F10 application. DNA was extracted and purified from ten separate pre-treatment and post-treatment samples using a DNeasy plant mini kit (Qiagen, Crawley, UK). Samples of pollen, honey and flowers were macerated in microfuge tubes containing extraction buffer using mini pestles (VWR International, Lutterworth, UK). Bees were left intact. The DNA samples were then used in

SID 5 (2/05) Page 18 of 33

PCR with the species F primers listed above. In order to check the specificity of the primers, negative controls were included consisting of DNA from isolates of species A, B and C (as described in 3.3.1), and also DNA from the entomopathogenic fungus Paecilomyces farinosus. This fungus was not used to examine primer specificity in 3.3.1, and only a small number of isolates of it were screened against varroa in the previous project, HH0813SHB. It was included in this experiment because: (a) it occurs in the same family as species D – F, and hence there is a possibility that DNA from this fungus could be amplified by primers developed for species F; (b) recent findings from one of our other projects, which is investigating the ecology of entomopathogenic fungi, indicates that this fungus occurs naturally in UK soils, and hence could be a contaminant in the environmental samples collected in this experiment. The species F primers were found to generate a product with DNA from P. farinosus (this suggests that P. farinosus is more closely related to species F than species A, B or C). A new pair of primers was designed after inclusion of P. farinosus ITS sequences in multiple sequence alignments described above. At the same time other primers internal to this new pair were designed so that nested PCR could be undertaken (Table 14). This technique utilises the PCR product generated by the first (external) pair of primers in a second round of PCR with internal primers. Nested PCR provides increased sensitivity coupled with improved specificity.

An experiment to determine the sensitivity of detection of first and second round amplification was carried out. The concentration of a DNA solution extracted from isolate F10 was determined using the Nanodrop ND-1000 (see 3.3.1 above). A dilution series of this DNA was used to generate products in first and second round (i.e. nested) PCR. The minimum concentration of DNA that gave a visible product when separated on an agarose gel was found to be 2pg (first round) and 0.02pg (second round). These values represent the DNA content of approximately 100 and 1 conidia respectively.

Each of the two sets of 50 DNA samples was amplified with the pair of external primers on three separate occasions and products were separated and visualised as described above. An aliquot (0.5μl) of the fluid from each of the 300 PCR reactions using the external primers was then amplified with the nested primers under the same conditions as before. These products were also separated and visualised. The results obtained from first and second round PCR with the pre- and post-F10 powder treated samples are given in Table 15 (values presented are the percentage of samples that gave a visible product).

Detection of second round PCR product in all pre-treatment samples except for honey indicates that species F is present naturally at very low levels on bees and B. napus flowers. It is likely that worker bees acquire fungal spores while foraging and bring them back into the bee colony, although the data indicates that these are not transferred into honey. The results also indicate that the first round PCR amplification gives the more meaningful indication of the movement of fungal inoculum following its application to the bee colony. Post treatment, species F was detected using first round amplification in all samples of foraging worker bees and brood nest adult workers, indicating that fungal inoculum applied to the colony successfully adhered to bees. Using first round amplification, the fungus was also detected on B. napus flowers, indicating that it was carried by foraging workers into the environment. The fungus was also detected in pollen cells (possibly carried on pollen loads on foragers in the colony when it was treated with fungus). Despite the fungus being applied directly into the colony, it was detected in only 20% of honey samples. It is more likely that most of this occurred by fungal conidia entering honey cells directly when the inoculum was applied, rather than by bees transferring conidia into honey cells. It will be important in future work to design a fungus application system that minimises the probability of conidia entering honey cells, although it should be noted from 3.3.2 that conidia do not survive for long in honey.

Summary Conidia placed in honey survived for no longer than 7 days at 20C. Conidia of species C died within hours

of being held in honey. Nucleotide sequence information of the rRNA gene repeat unit is available, together with specific forward

and reverse primers, for isolates of species A,B, C, and F. Species F is present naturally at low levels on adult bees. Species F conidia applied to the bee colony were detected on all foraging worker bees and brood nest

adult workers sampled. The fungus was also detected in pollen cells, but despite being applied directly into the colony, it was detected in only 20% of honey cells.

It will be important in future to design a fungus application system that minimises the probability of conidia entering honey cells

O4. (Objective 4): engagement with key actors Throughout the project we engaged with the bee keeping community at grassroots level (via presentations to local bee keeper associations, from Cornwall to Scotland; see Outputs section) and at national level, through dialogue with organisations such as the British Bee Keepers Association and Bee Disease Insurance. We have also been in dialogue with the biocontrol industry over the prospects for development of a commercial product. A key issue for commercialisation concerns product authorisation, and in particular the extent to which the biological data requirements for registration, as set by the regulator, act as a barrier to market entry.

SID 5 (2/05) Page 19 of 33

Products that are marketed in the EU for the prevention or treatment of disease in animals must be authorised under the terms of EC Directive 2004/28/FC, which is aimed at controlling veterinary medicines, including those that may enter the food chain, and provides regulations on safety, quality and traceability. In the UK, the Directive is enacted through the Veterinary Medicines Regulations 2005 and is regulated by the Veterinary Medicines Directorate (VMD). Medicinal products applied to A. mellifera fall within the remit of the Regulations and only approved products can be used by bee keepers. Representatives of UK bee keepers have expressed a view that, with the exception of antibiotics, bee medications should be treated as minor use products on minor species, in order to avoid a situation in which they could only be used by or obtained from a veterinary practitioner (Shaw, 2005). Insecticides (which presumably include any acaricide applied to bees for varroa control) require marketing authorisation as they are deemed to be medicinal in function (Veterinary Medicines Directorate, 2005, see http://www.vmd.gov.uk/General/VMR/vmr05.htm). Therefore, it is assumed that a marketing authorisation would also apply to a microbial control agent of varroa. Products applied to animal bedding and / or housing fall under the Biocidal Products Regulations administered by the Health and Safety Executive, but it is unlikely that this would apply to a varroa product applied to bee hives as the intention is to control mites feeding on bees.

It is important that the regulatory authority acts to ensure the safety of products but at the same time does not unduly inhibit their commercial development. The VMD declined our request for an overarching discussion on how biological control agents could fit in with the veterinary medicines registration system. However, they were willing to meet with us to discuss specific issues relating to the submission of a registration package once we had (a) secured an industrial partner to manufacture a finished product, and (b) appointed a pharmaceutical company or consultant experienced with working with dossiers for veterinary medicinal products. A similar experience with the VMD has been reported informally to the trades association, the International Biopesticide Manufacturers Association (IBMA), from a biopesticide manufacturer. The reluctance to hold pre-submission meeting is likely to act as a barrier to efforts to make the regulatory process fit-for-purpose. Specifically, it could unintentionally inhibit the commercialisation of a fungal control agent of varroa by putting into place inappropriate data requirements for authorisation. Factors contributing to regulatory failure are likely to include the following: (a) lack of knowledge by the regulator on biological control agents and the challenges facing their commercialisation; (b) the fact that the Veterinary Medicines Regulations were not designed with biological control agents in mind; (c) averseness to taking risks by the regulator (Hood et al., 2001). It should be remembered that market authorisation is going to be more costly for small to medium enterprises, which are less able to cope with high registration costs, but which have traditionally taken the lead in developing biopesticides. On the positive side, there is evidence that the regulatory system can adapted for new types of product. For example, provisions have been put in place by VMD for homeopathic veterinary medicines, which - like microbial control agents - are low risk products. Moreover, EC Directive 2004/28/FC states that ‘In certain circumstances … the need to obtain a marketing authorisation for a veterinary medicinal product in accordance with Community provisions is clearly disproportionate’, indicating that there is willingness at the EU level for the regulations to be adaptable as required. .

There are important parallels to be drawn here with the development and use of microbial biopesticides as plant protection products (PPPs) which, despite their perceived environmental and social benefits, have been taken up poorly in the UK due to socio-economic factors centred on regulation. The safety and efficacy of fungal, bacterial and viral biopesticides used as PPPs are regulated in the UK by the Pesticides Safety Directorate. The regulatory system used for these products was developed according to a chemical pesticides model, aspects of which have acted as barriers to biopesticide commercialisation (Advisory Committee on Pesticides, 2003). The outcome is that the entry costs to the PPP market for biopesticides were elevated disproportionately to the size of the market. A principal concern has been that ‘biopesticide development is locked into an inflexible and unimaginative chemical pesticide model. In this position, all of the shortcomings of biopesticides relative to chemicals emerge and none of the benefits’ (Waage, 1997). This should not be construed as a criticism of the regulators themselves and their day to day work, but rather it refers to ‘systemic problems that arise in regulation and which can be translated to the specific case of biopesticides’ (Grant, 2005). An unintended feature of regulatory systems is that specific and short-term goals can prevail over long-term aims (which in the case of PPPs concern the establishment of sustainable, environmentally benign systems of pest control). This is due to the bureaucratic tendency for policy instruments to be considered in isolation from their wider effects (Grant, 2005), and it can result in regulatory failure. In the case of biopesticides for plant protection, regulatory failure was recognised and addressed as a result of collaborative work between PSD and the Business Regulation Team (BRT) of the Regulatory Impact Unit of the Cabinet Office, starting in 2002. It was noted by the BRT that PSD’s authorisation requirements ‘were evidently designed to cope with standard, mass-produced synthetic chemical pesticides which, by their nature, tend to deliver very high efficacy rates, and not with this group of safer alternatives.’(Business Regulation Team, 2003, quoted from Grant, 2005). As a result of joint work between PSD and BRT, the former instigated a series of innovative measures including the introduction of a Pilot Scheme, pre-registration meetings with applicants, and dialogue with the IBMA. Feedback from the IBMA indicates that these initiatives have been very well received by the biopesticides industry. There is evidence that these changes have had an impact. At the time of writing, at least four biopesticides had been put forward for registration since the intervention by BRT.

SID 5 (2/05) Page 20 of 33

It is clear that the recent history of regulation of microbial biopesticides as PPPs could be used to inform the development of regulations for biopesticides intended for use as veterinary medicines. It is important to remember that, for all intents and purposes, we are dealing with identical products: They are based on the same organisms, with the same mode of action, and the same issues surrounding application and formulation technology and method of entry into the environment following application to the target. Writing about the future of biopesticides as components of Integrated Pest Management, Waage (1997) stated that ‘it is not the industry alone, but the entire pesticide regulatory process which has not adapted itself to the new opportunities which biopesticides provide.’ His comments may well have prescience for the regulation of biopesticides as veterinary products. However, there is an opportunity to learn from the experience of the Pesticides Safety Directorate and to develop a system of regulatory innovation.

Overall conclusions & options for new researchThe varroa mite continues to be a global threat to A. mellifera populations. Given the spread of chemical pesticide resistance in the UK and elsewhere, it is essential that new methods for its control are developed. Our research indicates that entomopathogenic fungi have potential as biological control agents of varroa. The results produced so far have been encouraging but there are still some issues to be resolved. For example, there are no candidate isolates that stand head and shoulder above the others, and hence isolate selection will come down to weighing up the costs and benefits of the various candidates. It must be borne in mind that the total number of isolates that we have been able to evaluate (c. 40 so far in all projects) is very small compared with the amount of genetic variation in virulence-related characteristics that exists within fungal species. Approximately 750 species of fungi in 56 genera are known to be pathogens or parasites of arthropods (Hawksworth et al., 1995), and we have examined just eight species in our work so far. This means that there are probably other, very desirable fungal isolates waiting to be identified. It must also be remembered that our screening programme is just a fraction of the size of the programmes run by agrochemical companies for the identification of new chemical actives from micro-organisms. We can make some generalisations based on the data collected from the isolates examined thus far as follows: Species F isolates are very desirable because they are pathogenic to varroa and have no effect on bees, but it would be of benefit if isolates could be identified with a higher optimum temperature for development. Species C isolates are often highly virulent to varroa, and in addition there are isolates with good temperature characteristics, and they are the easiest to mass produce (which can be a deal breaker in terms of profitable production of a commercial product). Unfortunately, they have poorer shelf life characteristics compared to other isolates, and they can cause mortality in adult and larval bees. However there are probably opportunities to identify ‘bee friendly’ species C isolates. Species A isolates are also virulent to varroa, temperature characteristics are acceptable, and although some isolates can have a negative effect on bees, there is strong evidence that species A conidia would be able to survive under bee colony environmental conditions considerably longer than other species, which could be critical for efficacy.

The development of a commercial product will require more research, although it should be remembered that the costs of developing a fungal biological agent are going to be markedly lower than the costs of developing a new chemical pesticide (which currently stands at about £150 million; the development costs of a fungal agent are likely to be around 1 – 5 % of this for a global market we estimate to be worth in excess of £30 million p.a.). The approach taken in this project has been not just to try to identify ‘winning’ fungal isolates for varroa control, but also to provide essential, underpinning knowledge for the development of an effective fungal-based control system. Indeed, the systems that we have developed could be used by others to develop alternative biological agents, should this be required.

Priorities for future research on fungal control of varroa should be as follows: (a) rapid screening to identify high temperature isolates of species D – F or bee –friendly species A & C isolates; (b) refinement of mass production and processing to improve product shelf life; (c) development of a more effective system for evaluating fungal efficacy; (d) development of an appropriate application system; (e) development of an isolate-specific detection method (to protect intellectual property); (f) evaluation of the phylogenetic relationships of candidate isolates with other isolates, especially those for which there is information published that is relevant to the biological data requirements for authorisation.

A central theme of this research has been to identify and overcome scientific and technical barriers to the development of new, sustainable methods of varroa control. However, in order to make sustainable varroa management a reality, it will be necessary to overcome regulatory barriers as well. As described above, recent developments concerning the regulation of microbial biopesticides as plant protection products could be used as a model for the development of an appropriate regulatory framework for biocontrol agents used as veterinary medicines. This will require the collaboration of a range of key actors from government, stakeholders and industry.

SID 5 (2/05) Page 21 of 33

Intellectual PropertyThe information generated by this project constitutes intellectual property that would be used to underpin and direct the development of a commercial product for varroa control. Because of the value of the information, we have not revealed the species or isolates used in the research. The information could be sold as a stand alone package to a commercial organisation, or it could form part of an intellectual property agreement in a strategic partnership with industry to develop a biocontrol agent. On the other hand, it may be appropriate to publish this information in peer review journals, and then use this as part of the submission to the biological data requirements for product registration. However, we have deliberately held back from publishing papers from this research until the most appropriate way forward has been decided.

AcknowledgementsThis research was sponsored by Defra. We are grateful to Norman Carreck (Rothamsted Research) for technical advice and assistance, and to Suzanne Clark (Rothamsted Research), Andrew Mead (Warwick HRI) and Carole Wright (Warwick HRI) for statistical analysis and advice on experimental design.

APPENDIX 1: Project outputs

Ball, B.V. (2002). Biological Control of Varroa. Presentation at the Scottish Beekeepers Association Autumn Convention, October 12th 2002.

Ball, B.V. (2002). Using fungi to fight the mite. Presentation at the Chalfont Beekeepers Association, Buckinghamshire, 26th November 2002.

Ball, B., Pell, J., Chandler, D., Davidson, G., Sunderland, K. & Carreck, N. (2002). Microbial control of varroa, phase II: realising the potential. Poster presentation at the National Honey Show, Kensington 14-16 th

November 2002.Ball, B., Pell, J., Chandler, D., Davidson, G., Sunderland, K. & Carreck, N. (2003). Microbial control of varroa,

phase two: realising the potential. Poster presentation at the BBKA Spring Convention, Stoneleigh, 26th April 2003.

Ball, B. V. (2004) Biological control of Varroa destructor. Presentation at Reading Beekeepers Association. February 12th 2004.

Ball, B.V. (2004) Fighting the mite with fungi. Oxfordshire Beekeepers Association Annual meeting. March 13 th

2004.Ball, B.V. (2004) Microbial control of Varroa destructor. Presentation at the Devon Beekeepers Association

residential conference, University of Exeter, July 16th-18th 2004.Ball, B.V. (2004) Microbial control of Varroa destructor. Presentation to Cheshire Beekeepers Association,

October 16th 2004.Ball, B V (2005) Using fungi to fight the mite. Presentation to Meon Valley Beekeepers Association, March 12 th

2005.Ball, B.V. (2005) Fighting the mite using fungi. BBKA Spring Convention, Stoneleigh, April 16th. Ball, B.V. (2005) Fungal biocontrol of Varroa destructor. Presentation to the executives of the national

beekeeping associations of England, Scotland, Wales, and Northern Ireland. February 23rd 2005.Ball, B.V. (2005) Microbial control of Varroa destructor. Presentation to Austrian Beekeepers, August 31st. Ball, B.V. (2005) Varroa control using fungi. Presentation to Sutton Coldfield Beekeepers Association, June 18 th

2005.Ball, B.V. Birchall, C., Carder, J., Chandler, D., Pell, J.K. & Prince, G. (2005). Dissemination and Persistence of

SID 5 (2/05) Page 22 of 33

Fungi to Control Varroa destructor. Apimondia Meeting Dublin, August 2005Birchall, C. (2004) Biological control of Varroa destructor. Presentation at Cambridgeshire Beekeepers’

Association One day Meeting 13th March 2004.Birchall, C. (2004) Biological control of Varroa destructor. 8th Rosena Clark Memorial Lecture, Twickenham

and Thames Valley Beekeepers’ Association 12th Nov 2004.Birchall, C., Davidson, G., Ball, B., Pell J. & Chandler D. (2004). Biological control of Varroa destructor. Poster

presentation at the BBKA Spring Convention, Stoneleigh, 24th April 2004.Birchall, C., Davidson, G., Ball, B., Pell J. & Chandler D. (2004). Biological Control of Varroa destructor -

dissemination and impact of spore inoculum. Proceedings of the 37 th Annual Meeting of the Society for Invertebrate Pathology, Helsinki, Finland, 1-6th August 2004.

Birchall, C., Pynson, B., Davidson, G., Ball, B., Pell, J., Chandler, D. (2004). Biological control of Varroa destructor – impact of spore inoculum on bees. Poster presentation at the COST Action 842 Meeting of the Working groups 1 and 2, U.K., 1-5th September, 2004.

Birchall, C., Davidson, G., Ball, B., Pell J. & Chandler D. (2004). Biological control of Varroa destructor. Proceedings of the European Conference of Apiculture, Udine, Italy, 19-23rd September 2004.

Birchall, C., Pynson, B., Davidson, G., Ball, B., Pell, J., Chandler, D. (2004). Biological Control of Varroa destructor - dissemination and impact of spore inoculum. Poster presentation at the 16 th Annual Meeting of the British Invertebrate Mycopathologists Group 27th October 2004.

Birchall, C., Davidson, G., Ball, B., Pell J. & Chandler D. (2004). Biological control of Varroa destructor. Poster presentation at the National Honey Show, Kensington 13-15th November 2003.

Birchall, C (2005) Biological control of Varroa destructor. Yorkshire Beekeepers’ Association Annual Conference, York. 12 Nov 2005

Birchall, C. (2005) Biological control of Varroa destructor – an update (2005). Presentation to Suffolk Beekeepers’ Association, Fornham St. Martin, Suffok. 17th Nov 2005

Birchall, C., Pynson, B., Davidson, G., Ball, B., Pell, J. & Chandler, D. (2005) Biological control of Varroa destructor – impact of spore inoculum on bees. Proceedings of the 39th International Apicultural Congress, Dublin 21-26 August.

Birchall, C (2006) Biological control of Varroa destructor. Bedfordshire Beekeepers’ Association Spring Meeting, Gamlingay, Bedfordshire. 19th April 2006.

Carreck, N.L. (2004) Biological control of varroa. Presentation to Bromley Beekeepers Association, January 24th.

Carreck, N.L. (2004) Fighting the mite. Presentation at Welsh Beekeepers Annual Convention, March.Carreck, N. L. (2004) Using fungi to fight the mite. Presentation at the Hampshire Beekeepers Association

Annual Conference October 9th. Carreck, N. & Birchall, C. (2005) An update on bee studies at Rothamsted Research. The Irish Beekeeper 60

(2) 41-45.Carreck, N. & Birchall, C. (2005) An update on bee studies at Rothamsted Research. The Welsh Beekeeper

148 4-18.Chandler, D. & Davidson, G. (2002). Biological control solutions come to bees’ rescue. HRI web page

( http://www.hri.ac.uk ) Chandler, D. & Davidson, G. (2002). Fungal control of bee killer. Appropriate Technology 29 (3), 22.Chandler, D. & Davidson, G. (2002). Biological Control of Varroa destructor. Presentation at the 14th meeting

of British Invertebrate Mycopathologists Group, Bath University, 18th September 2002.Chandler, D. & Davidson, G. (2002). Fungal control of Varroa destructor. The Welsh Beekeeper, 137, pp31-

33.Chandler, D. & Davidson, G. (2002). Fungi help combat honeybee killer. BBC news online

(http://news.bbc.co.uk/1/hi/sci/tech/2182948.stm).Chandler, D. (2002). Biological control solutions come to bees’ rescue. Presentation at launch of HRI annual

report, November 2002.Chandler, D. & Davidson, G. (2003). Fungal biocontrol of varroa. Presentation to the Barnet and District

Beekeepers Association, The Stableroom, Barnet, 22nd March 2003.Chandler, D. & Davidson, G. (2003). Hidden allies against varroa. Presentation to the Sutton Coldfield and

North Birmingham Beekeepers Association, The Methodist Church, Erdington, 16th January 2003.Chandler, D. & Ball, B. (2004). Fungal biocontrol of varroa. Presentation to annual Bee Inspectors meeting,

Central Science Laboratory, York, UK, 14-15th April 2004.Chandler, D. (2006). Fungal control of varroa . Raymond Ripley Memorial Lecture to West Cornwall Bee

Keepers Association, Duchy College, Cambourne, Cornwall, 4th March 2006Chandler, D. (2006). Fungal control of varroa. Presentation to Sutton Coldfield Bee Keepers Association, The

Methodist Church, Erdington, 20th April 2006. Davidson, G. & Chandler, D. (2002). Fungal control of Varroa destructor. Bee Craft, August 2002, pp24-26.Davidson, G. & Chandler, D. (2002). Fungal Control of Varroa Destructor. Bee Craft, August 2002, Pp24-26.Davidson, G. & Chandler, D. (2002). Fungal control of Varroa destructor. The BBKA newsletter, September

2002, No. 137.Davidson G., Phelps, K., Sunderland, K.D., Pell, J.K., Ball, B.V., Shaw, K.E. & Chandler, D. (2003). Study of

temperature-growth interactions of entomopathogenic fungi with potential for control of Varroa

SID 5 (2/05) Page 23 of 33

destructor (Acari: Mesostigmata) using a nonlinear model of poikilotherm development. Journal of Applied Microbiology, 94, 816-825.

Davidson, G., Birchall, C., Pell, J.K., Ball, B.V., Sunderland, K & Chandler, D. (2003). Evaluation of entomopathogenic fungi for control of Varroa destructor, an ectoparasite of the honey bee, Apis mellifera L. Proceedings of the 36th Annual Meeting of the Society for Invertebrate Pathology, Burlington, USA, 26-30 July 2003, p76.

Davidson, G. & Townsend, M. (2003). Fungal control of varroa. The BBKA newsletter, August 2003, No. 142.Davidson, G. (2003). Biological control comes to bees’ rescue. Presentation at Horticulture in Focus, Royal

Lancaster Hotel, London, 27th February 2003.Davidson, G., Birchall, C., Pell, J.K., Ball, B.V., Sunderland, K & Chandler, D. (2003). Evaluation Of

Entomopathogenic Fungi For Control Of Varroa Destructor, An Ectoparasite Of The Honey Bee, Apis Mellifera L. Proceedings Of The 36th Annual Meeting Of The Society For Invertebrate Pathology, Burlington, USA, 26-30 July 2003, P76.

Davidson, G., Birchall, C., Pell, J., Ball, B. & Chandler (2004). Evaluation of entomopathogenic fungi for control of Varroa destructor, an ectoparasite of the honey bee, Apis mellifera L. Proceedings of the British Mycological Society Conference “Fungi in the Environment”, Nottingham, UK, 13-15th September 2004.

Davidson, G., Birchall, C., Pell, J., Ball, B. & Chandler, D. (2004). Investigation of the survival of conidia of entomopathogenic fungi with potential for control of Varroa destructor in honey bee colonies. Proceedings of the 37th Annual Meeting of the Society for Invertebrate Pathology, Helsinki, Finland, 1-6th August 2004.

Davidson, G., Birchall, C., Pell, J., Ball, B. & Chandler, D. (2004). Physiological responses to temperature of potential fungal pathogens of Varroa destructor. Proceedings of the European Conference of Apiculture, Udine, Italy, 19-23rd September 2004

Davidson, G., Birchall, C., Pell, J., Ball, B. & Chandler, D. (2004). Hidden allies against varroa. Bee Craft, November 2004

Davidson, G., Birchall, C., Pell, J., Ball, B. & Chandler, D. (2004). Hidden allies against varroa. The Welsh Beekeeper, Winter 2004.

Pell, JK, Chandler, D., Birchall, C., Carder, J., Prince, G. and Ball, B.V. (2006) Fungal biocontrol of Varroa destructor, a major pest of the honey bee, Apis mellifera L. International Congress of Acarology, Amsterdam, August 2006

APPENDIX 2: Tables & Figures

Table 1: Isolates used in the study

Species Isolate Host or source

Country MTD (h) at 25ºC & 100% RH*

MTD (h) at 30ºC & 40% RH**

Optimum growth temperature (ºC)***

Super optimum growth temperature (ºC)****

A 1 Coleoptera USA 58 63 95

29.8 32.6

2 - - 60 - - -B 3 Acari Jamaica 91 304 27.7 39.6C 4 Acari USA 44 56 28.3 32.7

5 Acari - 58 60 26.4 35.7

6 Coleoptera - 57 74 26.6 34.27 - - 56 - 27.9 33.9

D 8 Homoptera UK 56 - - -E 9 Acari UK 56 84 23.3 32.7F 10 Homoptera UK 51 - 22.9 31.9

*Mean Time to Death measured in laboratory bioassays in previous research at 108 ml-1 except B3 (106 ml-1)(Shaw et al., 2002). Bioassays done with conidia suspensions at a concentration of 1 x 108 ml-1 except for B3 which was done at 1 x 106

ml-1 because of low spore yield with the species. ** Mean time to death measured in laboratory bioassays in previous research at two concentrations of conidia with selected isolates (Shaw et al., 2002).

Table 2: Virulence of single spore isolates to V. destructor in a laboratory bioassay

Rank Species Isolate Parent culture MTD* (h) s.e. % mortality at

SID 5 (2/05) Page 24 of 33

accession No. 7dpi**1 C 11 5 38.9 0.26 1002 C 12 6 39.7 1.07 1003 A 13 1 41.9 1.07 1004 C 14 4 41.9 5.39 1005 F 15 10 42.3 0.93 1006 E 16 9 42.8 2.05 1007 A 17 2 42.9 2.15 1008 C 18 7 43.8 3.47 1009 D 19 8 58.6 6.43 100

10 B 20 3 102.5 33.38 2511  Control 132.9 22.96 10

* Mean time to death**days post inoculation

Table 3: Average yields of aerial conidia of four fungal isolates produced on rice substrate, and conidia viability (% germination) at different stages in the production process

Isolate Log10 conidia g-1 dry weight substrate (s.d.)

Conidia g-1 dry weight substrate (back transformed)

Mean viability before pre-harvest drying (%)

Mean viability after pre-harvest drying (%)

Mean viability after harvest & processing (%)

A1 7.60 (0.256) 4.01 x 107 97 91 100B3 7.85 (0.275) 7.15 x 107 98 99 99F10 8.15 (0.181) 1.40 x 108 99 99 100C4 8.77 (0.299) 5.88 x 108 95 98 95Average moisture content (%) of the rice substrate before pre-harvest drying 19.8 (s.d. 6.27)Average moisture content (%) of the rice substrate after pre-harvest drying 3.9 (s.d. 2.80)

Table 4. Time (days) required for 50 % (VT50) and 5 % (VT5) of conidia to remain viable during long term storage at -20, +4 or +20ºC for three fungal isolates

Isolate Storage temperature ºC-20 4 20

VT50 (d) VT5 (d) VT50 (d) VT5 (d) VT50 (d) VT5 (d)A1 549

(361 – 1488)1525 (949 – 4497)

479 (329 – 1063)

1350 (877 – 3258)

202 (167 – 263)

539 (430 – 744)

C4 84(46 – 132)

588 (414 – 1090)

27( 11 – 42)

196 (153 – 279)

7 (1 – 13)

60 (44 – 98)

F10 416 (273 – 1205)

1393 (852 – 4551)

266 (203 – 411)

813 (590 – 1376)

140 (116 – 179)

405 (326 – 549)

Table 5: Germination of conidia of fungal isolates following exposure to 32.5C / 40% RH for a period of 39.5h

Species Isolate % germination (95% CL)A 1 70.6 (5.44)

2 73.1 (5.34)B 3 50.0 (12.21)C 4 51.6 (8.78)

5 52.4 (7.11)6 52.0 (11.78)7 55.2 (15.73)

D 8 62.3 (10.78)E 9 56.1 (13.63)F 10 58.4 (12.28)

SID 5 (2/05) Page 25 of 33

Table 6. IGT50 values estimated for fungal isolates following exposure to 32.5C / 40% RH for periods of between 0 and 168 h

Species Isolate IGT50 h (fiducial limits)A 1 163.9 (130.0 - 231.7)

2 183.0 (160.7 - 215.1)C 4 45.5 (40.1 - 51.3)

Table 7. Effect of carrier powders on mortality of newly emerged adult bees

Treatment % mortality standard error (n in parenthesis)

Control 13.1 3.4 (99)Bentonite 2.0 1.4 (99)Diatomaceous earth

7.1 2.6 (98)

Kaolin 10.1 3.0 (99)Talc 6.0 6.0 (100)

Table 8: Experiments to determine whether bees contaminated with F10 technical powder can indirectly initiate infection in mites placed on them: Mean percentage mite mortality in three experiments

MortalityExpt F10

powderconfidence interval (95%)

Heat -treated F10 powder

confidence interval (95%)

Control confidence interval (95%)

1 72 61-81 - - 34 24-462 58 44-71 12 5-25 15 7-303 53 35-70 22 11-39 23 11-42

Table 9: Percentage of mite cadavers that sporulated following treatment with F10 technical powder treatment (95% confidence intervals backtransformed from the logit scale).

Expt % mycosed (n) 95% confidence intervals1 82 (72) 49 - 952 77 (95) 53 - 903 83 (78) 46 - 97

Table 10: Treatment of bees with F10 technical powder: total percentage bee mortality averaged over all cages

Expt Percentage Mortality

SID 5 (2/05) Page 26 of 33

F10 powder

confidence interval (95%)

Heat treated F10 powder

confidence interval (95%)

Control confidence interval (95%)

1 0 - - 0 2 27 18 - 38 15 8 - 28 6 2 - 163 25 10 - 48 14 4 - 38 8 2 - 31

Table 11: Application of F10 technical powder to a nucleus colony: pattern of varroa mortality over the experimental period.

Time period MortalityNo. dead mites No. mycosed mites % mycosis

Pre 1st inoculation (19 days)

263 (18 day-1)

-

Post 1st inoculation (18 days)

194(13 day-1)

104 53.6

Pre 2nd inoculation (50 days)

328(7 day-1)

-

Post 2nd inoculation (19 days)

329(17 day-1)

176 53.5

Total 1114

Table 12. rRNA ITS forward and reverse primers designed and evaluated for species used in the project

Primer sequence TmSpecies A forward primer 2 gacgcggactggaccagcg 66 Species A reverse primer 1 cggtgcgagctgtattactg 62

Species B forward primer 2 ctcttgtatctggatgcattgc 60 Species B reverse primer 2 cggcggactcgtcctcc 60

Species C forward primer 1 ccaacccctgtgaattatacc 58 Species C reverse primer 1 cgatccccaacaccaagtc 61

Species F forward primer 1 gtccggacggcctcgc 58 Species F reverse primer 1 ggttccggtgcgagttgg 60

Table 13. Survival of fungal conidia stored in honey at different temperatures. Survival expressed as percentage of the number of colony forming units in controls (0.05% Triton X 100 used in place of honey with 0 days incubation)

Species Isolate Temperature % survival (s.e.m.)(C) Day 0 Day 7 Day 14

A 1 20 8.9 (1.39) 0 05 13.1 (6.11) 0 0

F 10 20 76.1 (8.18) 0 05 77.7 (3.39) 6.5 (0.94) 0

C 4 20 0.4 (0.22) 0 0

SID 5 (2/05) Page 27 of 33

5 0.4 (0.44) 0 0

Table 14. Forward and reverse primers designed for nested PCR amplification of the rRNA ITS region of species F

Primer sequence TmSpecies F forward primer 2 cccttatgtgaacatacctact 62Species F reverse primer 2 gtgttttacggcgaggcca 60

Species F nested forward primer 1 gcccgcggcccggac 58 Species F nested reverse primer 1 ccgatttccccaaagggaag 62

Table 15: Detection of species F in environmental samples following application of the fungus to a bee colony, using nested PCR amplification of rRNA ITS. Values presented are the percentage of samples that gave a visible product.

Type of sample Pre-treatment Post-treatment1st round PCR 2nd round PCR 1st round PCR 2nd round PCR

Live foragers 0 90 100 100Brood nest workers 0 70 100 100B. napus flowers 0 90 100 100Pollen 0 90 90 100Honey 0 0 20 90

Figure 1: Proportion of capped brood after larval feeding with entomopathogenic fungi

Figure 2: Total daily mite mortality and mite mortality due to mycosis

SID 5 (2/05) Page 28 of 33

0.20.0

0.4

0.8

1.0

Spe

cies

F

Spe

cies

C

cont

rol,

no fu

ngus

pre

sent

cont

rol,

F pr

esen

t

cont

rol,

C p

rese

nt

Twee

n, n

o fu

ngus

pre

sent

Twee

n, F

pre

sent

Prop. capped cells

Twee

n, C

pre

sent

Figure 3: Mycosis of mites after inoculation

SID 5 (2/05) Page 29 of 33

0

5

10

15

20

25

30

35

03/11/2005 10/11/2005 17/11/2005 24/11/2005 01/12/2005 08/12/2005 15/12/2005 22/12/2005 29/12/2005 05/01/2006 12/01/2006 19/01/2006 26/01/2006 02/02/2006 09/02/2006 16/02/2006

NUMBER OF DEAD MITES

Total mite mortality

Mite mortality due to mycosis

DATE

APPENDIX 3: Reference to published material

Advisory Committee on Pesticides (2003). Alternatives to conventional pest control techniques in the UK: A scoping study of the potential for their wider use. http://www.pesticides.gov.uk/uploadedfiles/Web_Assets/ACP/ACP_alternatives_web_subgrp_report.pdf

Allen-Wardell, G., Bernhardt, P., Bitner, R., Burques, A., Buchmann, S., Cane, J., Cox, P.A., Dalton, V., Feinsinger, P., Ingram, M., Inouye, D., Jones, C. E., Kennedy, K., Kevan, P., Koopowitz, H., Medellin, R., Medellin-Morales, S. & Nabham, G. P. (1998). The potential consequences of pollinator declines on the conservation of biodiversity and stability of food crop yields. Conservation Biology 12, 8 – 17.

Andersen, M., Magan, N., Mead, A., & Chandler, D. (2006). Development of a population-based threshold model of conidial germination for analysing the effects of physiological manipulation on the stress tolerance and infectivity of insect pathogenic fungi. Environmental Microbiology (in press).

Bailey, L. (1971). The safety of pest-insect pathogens for beneficial insects. In, Burges, H. D., Hussey, N. D. (eds) Microbial Control of Insects and Mites. Academic Press, London, pp. 491 – 505.

Ball, B.V. (1993). The damaging effects of Varroa jacobsoni infestation, in Living with Varroa (Matheson, A., Ed.) International Bee Research Association, Cardiff, UK, pp. 9-16.

Ball, B.V. (1994a). The development of control strategies for Varroa jacobsoni in colonies of Apis mellifera. Proceedings of the Brighton Crop Protection Conference - Pests & Diseases 1994 2, 569-576.

Ball, B.V. (1994b). Host-parasite-pathogen interactions, in New Perspectives on Varroa, (Matheson, A., Ed.) International Bee Research Association, Cardiff, UK, pp. 5-11.

Bartlett M.C. & Jaronski, S.T. (1988). Mass production of entomogenous fungi for biological control of insects. In: Fungi in Biological Control Systems, ed. M.N Burges, pp 61-85. Manchester University Press, UK.

SID 5 (2/05) Page 30 of 33

Beetsma, J. (1994). The Varroa mite, a devastating parasite of western honeybees and an economic threat to beekeeping. Outlook on Agriculture 23, 169-175.

Bew, M. (1992). Varroasis disease of honeybees - diagnosis and control. MAFF, Luddington, UK.Bidochka, M.J., Pfeifer, T.A. & Khachatourians, G.G. (1987). Development of the entomopathogenic fungus

Beauveria bassiana in liquid cultures. Mycopathologia 99, 77-83.Bowen-Walker, P. L., Martin, S. J. & Gunn, A. (1999). The transmission of deformed wing virus between

honeybees (Apis mellifera) by the ectoparasitic mite Varroa jacobsoni Oud. Journal of Invertebrate Pathology 73, 101 – 106.

Buchler, R. (1994). Varroa tolerance in honey bees - occurrence, characters and breeding, in New Perspectives on Varroa (Matheson, A., Ed.) International Bee Research Association, Cardiff, UK, pp. 12-23.

Burges, H.D. (1998). Formulation of mycoinsecticides. In: Formulation of Microbial Biopesticides, ed. H.D. Burges, pp 131-185. Kluwer Academic Publishers, The Netherlands.

Chandler, D. (1994) Cryopreservation of fungal spores using porous beads. Mycological Research 98, 525-526.

Chandler, D., Davidson, G. & Jacobson, R. J. (2005). Laboratory and glasshouse evaluation of entomopathogenic fungi against the two-spotted spider mite, Tetranychus urticae (Acari: Tetranychidae) on tomato, Lycopersicon esculentum. Biocontrol Science and Technology 15, 37 – 54.

Chandler, D., Sunderland, K. D., Ball, B.V. & Davidson, G. D. (2001). Prospective biological control agents of Varroa destructor n. sp., an important pest of the European honeybee, Apis mellifera. Biocontrol Science and Technology, 11, 429 - 448.

Clerk, G.C. & Madelin, M.F. (1965). The longevity of conidia of three insect-parasitising hyphomycetes. Transactions of the British Mycological Society 48, 193-209.

Couch, T.L. & Ignoffo, C.M. (1981). Formulation of insect pathogens. In: Microbial Control of Pests and Plant Diseases 1970-1980, ed H.D. Burges, pp621-634. Academic Press, UK

Danka, R. G., Rinderer, T. E., Kuznetsov, V. N. & Delatte, G. T. (1995). A USDA-ARS project to evaluate resistance to Varroa jacobsoni by honey-bees of far-eastern Russia. American Bee Journal 135, 746-748.

Davidson, G. Phelps, K., Sunderland, K.D., Pell, J. K., Ball, B. V., Shaw, K. E., Chandler, D. (2003). Study of temperature-growth interactions of entomopathogenic fungi with potential for control of Varroa destructor using a nonlinear model of poikilotherm development. Journal of Applied Microbiology, 94, 816-825.

De Jong, D., Morse, R.A. & Eickwort, G.C. (1982). Mite pests of honeybees. Annual Review of Entomology 27, 229-252.

Delfinado-Baker, M. & Houck, M.A. (1989). Geographic variation in Varroa jacobsoni (Acarina, Varroidae): application of multivariate morphometric techniques. Apidologie 20, 345-358.

Donze, G. & Guerin, P.M. (1994). Behavioral attributes and parental care of Varroa mites parasitising honeybee brood. Behavioral Ecology and Sociobiology 34, 305-319.

Dorta, B., Bosch, A., Arcas, J.A. & Ertola, R.J. (1990). High level of sporulation of Metarhizium anisopliae in a medium containing by-products. Applied Microbiology & Biotechnology 33, 712-715.

Elzen, P. J., Eischen, F. A., Baxter, J. B., Pettis, J., Elzen, G. W. & Wilson, W. T. (1998). Fluvalinate resistance in Varroa jacobsoni from several geographic locations. American Bee Journal 138, 674 – 676.

Engels, W. (1994). Varroa control by hyperthermia, in New Perspectives on Varroa (Matheson, A. Ed.) International Bee Research Association, Cardiff, UK, pp. 115-119.

Feng, M.G., Poprawski, T.J. & Khachatourians, G.G. (1994). Production, formulation and application of the entomopathogenic fungus Beauveria bassiana for insect control: current status. Biocontrol Science and Technology 4, 3-34.

Fries, I. (1993). Varroa in cold climates, in Living with Varroa (Matheson, A., Ed.) International Bee Research Association, Cardiff, UK, pp. 37-48.

Fries, I. (1997). Organic control of Varroa, in Varroa: Fight the Mite (Munn, P. & Jones, R., Eds.) International Bee Research Association, Cardiff, UK, pp. 16-21.

Genstat (2000) Genstat for Windows. Release 4.2. Fifth Edition. VSN International Ltd. Oxford, UK.Goettel, M. S. and G. D. Inglis (1997) Fungi: Hyphomycetes, pp. 213 - 249. In, Manual of techniques in insect

pathology, ed. L. Lacey. Academic Press, San Diego, USA. Grant, W. P. (2005). The Challenges of Interdisciplinary Research: The Case of Biopesticides. Northeastern

Political Science Association Environmental Conference, Philadelphia, Pa. USA, November 2005. http://www2.warwick. ac.uk/fac/soc/pais/biopesticides/papers/

Harbo, J. R. & Hoopingarner, R. A. (1997). Honey bees (Hymenoptera: Apidae) in the United States that express resistance to Varroa jacobsoni (Mesostigmata: Varroidae). Journal of Economic Entomology 90, 893 – 898.

Hawksworth, D. L. (1991). The fungal dimension of biodiversity: magnitude, significance and conservation. Mycological Research 95, 641 – 655.

Hillesheim, E., Ritter, W. & Bassand, D. (1996). First data on resistance mechanisms of Varroa jacobsoni (Oud.) against tau-fluvalinate. Experimental and Applied Acarology 20, 283 – 296.

Hong, T.D., Jenkins, N.E. & Ellis, R.H. (2000). The effects of duration of development and drying regime on

SID 5 (2/05) Page 31 of 33

the longevity of conidia of Metarhizium flavoviride. Mycological Research 6, 662-665.Hood, C., Rothstein, H. & Baldwin, R. (2001). The Government of Risk: Understanding Risk Regulation

Regimes. Oxford University Press, Oxford, UK, 228pp. Jaronski, S.T. (1997). New paradigms in formulating mycoinsecticides. In: Pesticide Formulations and

Application Systems, eds G.R. Goss, M.J. Hopkinson and H.M. Collins, pp 99-112. American Society for Testing and Materials 17, USA.

Jenkins, N.E. & Lomer, C.J. (1994). Development of a new procedure for the mass production of conidia of Metarhizium flavoviride. IOBC/SROP 17(3) pp181-184.

Jenkins, N.E. & Thomas, M.B. (1996). Effect of formulation and application method of the efficacy of aerial and submerged conidia of Metarhizium flavoviride for locust and grasshopper control. Pesticide Science 46, 299-306.

Kanga, L. H. B., Jones, W. A. & James, R. R. (2003). Field trials using the fungal pathogen Metarhizium anisopliae (Deuteromycetes: Hyphomycetes) to control the ectoparasitic mite Varroa destructor (Acari: Varroidae) in honey bee, Apis mellifera (Hymenoptera: Apidae) colonies. Journal of Economic Entomology 96, 1091 – 1099.

Kleepsies, R.G. & Zimmermann, G. (1998). Effect of additives on the production, viability and virulence of blastospores of Metarhizium anisopliae. Biocontrol Science and Technology 8, 207-214.

Martin, S. J. (1997).The life and death of Varroa, in Varroa: Fight the Mite (Munn. P. & Jones, R., Eds.) International Bee Research Association, Cardiff, UK, pp. 3 – 10.

Matheson, A. (1994). The impact of Varroa infestation on beekeeping, in New Perspectives on Varroa (Matheson, A., Ed.) International Bee Research Association, Cardiff, UK, pp. 27 – 31.

Mendonca A.F. (1992) Mass production, application and formulation of Metarhizium anisopliae for control of sugarcane froghopper, Mahanarva posticata, in Brazil. In: Biological Control of Locusts and grasshopper, eds. C.J. Lomer and C. Prior, pp 239-244. Cabi International, UK.

Milani, N. (1994). Analytical bibliography on Varroa jacobsoni Oud. and related species. Apicoltura 8, 1-147.Milani, N. (1999). The resistance of Varroa jacobsoni Oud. to acaricides. Apidologie 30, 229 – 234. Mobus, B. & De-Bruyn, C. (1993). The New Varroa Handbook. Northern Bee Books, Hebden Bridge, UK.Moore, D. & Caudwell, R.W. (1997). Formulation of entomopathogens for the control of grasshoppers and

locusts. Memoirs of the Entomological Society of Canada 171, 49-67.Moore, D. & Higgins, P. M. (1997). Viability of stored conidia of Metarhizium flavoviride Gams and Rozsypal,

produced under differing culture regimes and stored with clays. Biocontrol Science and Technology 7, 335 – 343.

Moore, D. & Higgins, P.M. (1997). Viability of stored conidia of Metarhizium flavoviride Gams and Rozsypal, produced under differing culture regimes and stored with clays. Biocontrol Science and Technology 7, 335-343.

Moore, D., Douro-Kpindou, O.K., Jenkins, N.E. & Lomer, C.J. (1996). Effects of moisture content and temperature on storage of Metarhizium flavoviride conidia. Biocontrol Science and Technology 6, 51-61.

Oldroyd, B. (1999). Coevolution while you wait: Varroa jacobsoni, a new parasite of western honeybees. Trends in Ecology and Evolution 14, 312 – 315.

Ramirez, B.W. (1987). Biological analogies between some fig-wasps (Hymenoptera: Agaonidae and Torymidae: Sycophaginae) and Varroa jacobsoni (Acari: Varroidae). Revista de Biologia Tropical 35, 209-214.

Ritter, W. (1993). What future for biotechnical control ?, in Living with Varroa (Matheson, A., Ed.) International Bee Research Association, Cardiff, UK, pp. 25-31.

Sammataro, D., Gerson, U. & Needham, G. (2000). Parasitic mites of honey bees: life history, implications, and impact. Annual Review of Entomology 45, 519 – 548.

Sanford, T. (1996, June). World winter varroa kill. APIS (Apicultural Information and Issues from IFAS / University of Florida Department of Entomology and Nematology) 14 [WWW document]. URL http://www.ifas.ufl.edu/~mts/apishtm/apis96/apjun96.htm#8

Sanford, T. (1999, January). Raising the bar on pesticide use: Coumaphos approved in Florida. APIS (Apicultural Information and Issues from IFAS / University of Florida Department of Entomology and Nematology) 17 [WWW document] URL http://www.ifas.ufl.edu/~mts/apishtm/apis99/apjan99.htm#1

Shaw, K. E., Davidson, G., Clark, S. J., Ball, B. V., Pell, J. K., Chandler, D. & Sunderland, K. (2002). Laboratory bioassays to assess the pathogenicity of Mitosporic fungi to Varroa destructor (sp. nov) (Acari: Mesostigmata), an ectoparasitic mite of the honeybee, Apis mellifera L. Biological Control 24, 266 - 276.

Shaw, M. W. (2005). The veterinary medicines regulations 2005 : response on behalf of the Welsh Beekeepers Association. The Welsh Beekeeper 148, 26 – 27.

Thomas, H.U. (1997). Practical aspects of alternative Varroa control methods, in Varroa: fight the mite (Munn, P. & Jones, R., Eds.) International Bee Research Association, Cardiff, UK, pp. 22-30.

Waage, J. K. (1997). Biopesticides at the crossroads: IPM products or chemical clones? 1997 BCPC Symposium Proceedings no. 68: Microbial insecticides: novelty or necessity? British Crop Protection

SID 5 (2/05) Page 32 of 33

Council, Farnham, UK, p 11 – 20. Walton, S. (1996). SOS for British bees. The Grower, 26 September 1996, 11-12.White, T. J., Bruns, T., Lee, S. and Taylor, J. (1990). Amplification and direct sequencing of fungal ribosomal RNA

genes for phylogenetics. In: Innis, M. A., Gelfand, D. H., Sninsky, J. J. & White, T. J. eds. PCR Protocols. A guide to methods and application. San Diego, CA, USA: Academic Press, 315-322.

Wraight, S.A., Jackson, M.A. & de Kock, S.L. (2001). Production, stabilization and formulation of fungal biocontrol agents. In: Fungi as Biocontrol Agents, eds T.M. Butt, C. Jackson and N, Magan, pp , CAB International, UK.

References to published material9. This section should be used to record links (hypertext links where possible) or references to other

published material generated by, or relating to this project.

SID 5 (2/05) Page 33 of 33