Entomopathogenic fungi against whiteflies : tritrophic interactions between Aschersonia species, Trialeurodes vaporariorum and Bemisia argentifolii, and glasshouse crops

Many horticultural and agricultural crops are good host plants for the greenhouse whitefly, Trialeurodes vaporariorum, and the silverleaf whitefly, Bemisia argentifolii . Their damage to crops is manifold. When present in sufficient numbers they can cause leaf drop and inhibit fruit maturation. They are efficient vectors of economically important plant viruses. In addition, whiteflies produce honeydew, which soils and damages crops, and serves as a substrate for sooty moulds, thus reducing leaf photosynthesis and renders plants and fruits unsightly.

As more environmentally responsible agricultural strategies are adopted, natural enemies of both whitefly species, will play an increasing role. Screening for natural enemies which are able to kill both pest insects quickly, without affecting other natural enemies, is an important line of research. Entomopathogenic fungi can meet these requirements and can therefore be a valuable asset to existing biological and chemical control measures. Our attention is directed towards the entomopathogenic fungi of the genus Aschersonia , which are specialised on whitefly and scale insects. Previous research indicated that Aschersonia aleyrodis is a promising whitefly control agent because of its tolerance to relative humidities as low as 50%, its long persistence on leaf surfaces and its compatibility with the parasitoid Encarsia formosa , but little is known about other species within the genus.

This project consisted of two components: 1) to identify virulent isolates of Aschersonia spp. for the use against greenhouse and silverleaf whitefly, and 2) to study factors which influence the effectivity of entomopathogenic fungi, with special reference to host plant, humidity and their interaction.

Forty-four isolates of Aschersonia spp. were tested for their ability to sporulate on semi-artificial media and to infect the insect hosts, both important criteria for selection of biocontrol agents. Seven isolates sporulated poorly (less than 5.10 ⁷ conidia/culture) and ten were not able to infect either of the whitefly species. After a selection based on spore production and infection, virulence of 31 isolates was evaluated on third instar nymphs of both whitefly species on poinsettia ( Euphorbia pulcherrima ). Infection levels varied between 2 to 70%, and infection percentages of B. argentifolii correlated with that of T. vaporariorum . Several isolates, among which unidentified species of Aschersonia originating from Thailand and Malaysia, A. aleyrodis from Colombia, and A. placenta from India showed consistent results in their ability to infect both whitefly species. Of these isolates LD ₅₀ and LT ₅₀ values were compared to select the most virulent isolate for control of B. argentifolii and T. vaporariorum . Bioassays were carried out on intact poinsettia plants under glasshouse conditions and the dosage ranged from 14 to 1.4 x 10 ⁵ conidia/cm ² . On B. argentifolii LD ₅₀ 's varied between 1600 and 4800 conidia/cm ² and LT ₅₀ 's between 4.6 - 8.7 days and on T. vaporariorum LD ₅₀ 's varied between 700 and 5300 conidia/cm ² and LT ₅₀ 's between 4.5 and 9.9 days. For two isolates the optimum mortality did not occur at the highest dosage, which was also reflected in the speed of kill.

To obtain better insight into the infection process of A. aleyrodis , A. placenta and an unidentified Aschersonia sp. scanning electron microscopy and bioassays were carried out. Conidia of Aschersonia spp. germinated readily on the cuticula of host insects as well as on water agar. On water agar A. placenta also produced capilliconidia. No germination was observed on poinsettia leaf surface, except on the leaf veins. On B. argentifolii the fungi formed large amounts of mucilage to attach themselves to the insect. Appressoria were formed before penetration, but also direct penetration was observed. This seemed not related to a specific site on the insect. For both whitefly species, first to third instar nymphs were most susceptible. If the population existed of fourth instar nymphs for more than 50%, infection levels dropped from 90 to 50%. Infected whitefly nymphs usually died in the stage following the treated stage. The fungus protruded from the insect via the margins or via the emergence folds of pupae, if humidity levels were high enough.

Persistence of A. aleyrodis was studied on cucumber ( Cucumis sativus ), gerbera ( Gerbera jamesonii ) and poinsettia. Germination capacity and infectivity of conidia, which stayed on the different plants over a period of up to one month, were assessed. Average germination of conidia on the leaves was low (< 14%), whereas it was shown that most of the conidia transferred from the leaf to water agar were viable, even after having been present on the leaf surface for one month. Germination capacity was influenced by host plant species: it was highest on cucumber, followed by poinsettia and lowest on gerbera. On cucumber leaves, conidia stayed viable and were able to infect 90% of the whitefly nymphs, even at 31 days after spore application. On gerbera, germination capacity decreased considerably from 80% (day 0) to 40% (day 31). This was reflected in nymphal mortality, which declined from 75% to 40%. Despite the high germination capacity (60%) of conidia on poinsettia after a one month on the leaf surface, nymphal mortality decreased from 70% at the day of spore application to 10% after three days at leaf surface, and remained low throughout the monitoring period. The phyllosphere microflora, secondary plant metabolites and microclimate can play a role in these findings.

Can phyllosphere humidity explain differences in insect mortality due to entomopathogenic fungi on different crops? This was tested for cucumber, gerbera, tomato ( Lycopersicon esculentum ) and poinsettia on which mycosis of T. vaporariorum and cotton aphid ( Aphis gossypii ) were determined in relation to host-plant climate. Phyllosphere humidity was estimated using climate and host-plant parameters. Hydrophobicity of the leaves and crop density were also taken into account. On cucumber, gerbera and tomato, the fungi A. aleyrodis and A. placenta caused over 90% mortality in whitefly, while another entomopathogenic fungus, Verticillium lecanii , caused 50% mortality. On poinsettia, whitefly mortality was significantly lower for all three fungi (ca. 20%). The phyllosphere of cucumber was more humid than that of the other crops. This cannot the lower whitefly mortality on poinsettia. The fact that poinsettia leaves were more hydrophobic than the other leaves may offer an explanation for the observed lower whitefly mortality, but chemical host-plant aspects may also play a role. Our results underline the importance of the first trophic level (plant) for entomopathogenic fungi in integrated pest management programs.

The influence of relative humidity (RH) and host-plant species and their interaction on mycoses of T. vaporariorum by A. aleyrodis , A. placenta and V. lecanii was studied. Experiments conducted on poinsettia, gerbera and cucumber at 50% and 80% RH showed a clear host-plant effect. On cucumber and gerbera all fungi caused significantly higher mortality compared with the control and a higher humidity resulted in higher mortality by the fungus, whereas whitefly mortality on poinsettia remained low and was not significantly different from the control. Both Aschersonia spp. performed significantly better than V. lecanii . In following experiments an additional period of 0, 3, 6, 12, 24 or 48 hours of high RH was applied after treatment with A. aleyrodis and V. lecanii . Mortality caused by the fungus increased with a longer additional period of additional high RH and higher ambient humidity. However, the host-plant species effect exceeded the effect of ambient or additional high RH period; whitefly mortality was highest on cucumber and lowest on poinsettia. On poinsettia both fungi had hardly any effect on whitefly mortality, except after a 48 hrs additional high RH period. On gerbera and cucumber all fungal treatments were significantly different from the control and A. aleyrodis performed better than V. lecanii .

The influence of cultivar differences on the efficacy of entomopathogenic fungi is also tested. The experiment was conducted in a glasshouse (semi-practice) using two gerbera cultivars, 'Bourgogne' and 'Bianca' with different plant structure and trichome density on the leaf: 'Bianca' having more and larger leaves and fewer trichomes/cm ² than 'Bourgogne'. A.   aleyrodis and V. lecanii were applied for the control of T. vaporariorum , in two concentrations, 10 ⁶ and 10 ⁷ conidia/ml. In 'Bianca', both fungi caused a whitefly mortality up to 80%. Whitefly mortality was higher for 10 ⁷ than for 10 ⁶ conidia/ml. In 'Bourgogne', V. lecanii caused a significantly lower mortality than A. aleyrodis . Although no cultivar differences in whitefly development time were found, other characteristics, like natural mortality and build-up of the whitefly population, seemed to differ. Differences in mortality levels in relation to cultivar (humidity, hit-probability) and whitefly characteristics may explain the differences between the efficacy of the fungal species.

The host plant is the most important factor in this tritrophic system, exceeding the influence of ambient humidity. However, although entomopathogenic fungi are less successful in controlling whiteflies on poinsettia, these problems may be circumvented by use of formulation. Aschersonia spp. are highly virulent against whitefly pests, can cause natural epizootics in the field, can be grown on artificial media, are well adapted to survive in the canopy environment, and are compatible to/or even complementary with other natural enemies.