Evaluation location: Mississippi, United States
Basic ecology of host blueberry species and genotypes.
The genus Vaccinium L. (blueberries, cranberries, bilberries) in the Ericaceae (Heath family) includes upwards of 450 species, most inhabiting cooler tropical microclimates (e.g., forest understory and higher elevations). The remaining species occur throughout subtropical, temperate, and boreal regions of the northern hemisphere including regions native and exotic to D. suzukii. Insular species that we tested include V. reticulatum, V. calycinum, V. cylindraceum and V. padifolium, species hypothesized to be highly susceptible to attack. Blueberries and cranberries are currently the most economically important of North American Vaccinium. Cultivated highbush-type blueberries are polyploid, more specifically, tetraploid like northern and southern highbush blueberries (V. corymbosum) and hexaploid like rabbiteye blueberry (V. virgatum syn. ashei, Hummer et al. 2015, Kron et al. 2002). Of the 29 species tested, 4 Asian blueberry species, V. myrtoides, V. bracteatum, V. oldhami, and V. smallii, probably shared a long evolutionary history with D. suzukii and, hence, could have evolved some level of resistance.
Greenhouse growing conditions and germplasm handling.
The USDA National Clonal Germplasm Repository in Corvallis, Oregon conserves wild relatives, cultivars and hybrids of Vaccinium. Non-hardy types are maintained as containerized plants in greenhouses, whereas cold hardy types grow in the field. Fruit for these species ripen between June and August. For this study, samples representing 29 species including cultivars and hybrids (Table 1), were harvested at early to peak ripeness, and were shipped overnight, for analysis at the USDA-ARS facilities in Poplarville, Mississippi. Fruit of rabbiteye and southern highbush blueberry were collected inhouse from greenhouse-grown fruit at the Poplarville Location. About 30 chilled fresh fruit were sent for each of the 140 genotypes. As the 2016 season progressed and ripening fruit became available, six separate packages of berry samples were sent on 1 June, 21 June, 12 July, 20 July, 25 July, and 3 August. Information on the 29 species used for this study are shown in Table 1. Tables 2 through 5 also provide the genotypes that provided enough intact berries for resistance screening.
Fly cultures and handling.
Drosophila suzukii adults used in these tests originated from inbred cultures reared from flies originally collected in 2014 from rabbiteye blueberry (Vaccinium virgatum syn. ashei) fruit harvested at an experimental farm near Old Creek and Indian Camp Branch, Perkinston MS (30°47014″N, 88°59026″W). Fly cultures were maintained in 68 ml plastic vials. Our D. suzukii diet included the four ingredients: 3.75 g of standard Drosophila instant media (formula 4-24, Carolina Biological Supply, Burlington, NC), an intact sterile blueberry or blackberry fruit, 20 ml water, and 0.8 mg Fleishmann’s dry baker’s yeast added as a supplemental protein source (Sampson et al. 2016). Cultures developed under ambient indoor lighting at 21°C under 60% relative humidity.
Bioassay (Screening) Protocols.
We assessed resistance (antibiosis) using detached fruit bioassays. We recorded the resulting reproductive parameters of 12.0 ± 0.4 females caged with an equal number of males (11 ± 0.5 males, Mean ± SE, n = 229 vials). Flies within each vial had access to ~10 fruits or 15g of intact unblemished fruit containing their pedicels. Pedicels were important to retain because any open wound or incision on a fruit could permit flies to circumvent resistant surfaces and accelerate oviposition 2- to 15-fold (Ioriatti et al. 2015, Rezazadeh et al. 2018). Flies and fruit from randomly selected cultures were confined to randomly-chosen (68 ml) plastic Drosophila rearing vials solely containing fruit of a known berry species and genotype. Vials remained horizontal to give flies uniform access to berries for 3 days. Our bioassay should provide adequate pest pressure to reveal antibiosis because of a mean density of 1.2 females and 1.1 males per host berry. Flies were given no supplemental food because an ability to puncture fruit for feeding and laying eggs was considered an element of resistance. Initially, fly mortality went unassessed because adults were expected to survive a 3-day confinement period without food or water. However, when we discovered that adults must feed from fruit punctures to survive, we began assessing fly mortality on subsequent fruit from 46 other host genotypes, including genotypes designated as our non-host standard. After 3 days, dead flies of both sexes were counted and remaining live flies euthanized. Fruit were then analysed for mass (g), density (berries per vial), and brix or the percentage dissolved solids, a measure of sugar content or ripeness determined using a handheld refractometer. We simultaneously measured the following fly life history parameters: (p1) subsurface eggs (viable eggs) per fruit, (p2) surface eggs (unviable eggs) per fruit, (p3) larvae per fruit, (p4) adult females per vial, (p5) adult males per vial, (p6) adult mortality (dead flies / total flies *100), (p7) viable brood per female [(p1+p3)/p4], (p8) reproductive rate over 3 d = (p7 / fruit density).
A resistance index was calculated as the sum of the number of fly reproductive parameters not significant (indicated by **) from the host (Vaccinium) standard)) + 1 (for significance from non-host standard) or +2 (for non-significance from non-host standard). The minimal resistance value is 1, the maximum is 5.
The non-host resistance standard was conceived ad hoc and included fly reproductive responses to fruit collected locally from known or probable non-hosts such as Vitis rotundifolia, Ilex, Citrus, and Rhaphiolepsis. The host (Vaccinium) resistance standard was developed post hoc and included fly reproductive responses to highly-resistant host fruit collected from Vaccinium consanguineum, V. cylindraceum, V. ovatum, V. myrtoides, V. corymbosum cvs. CVAC1312 and ‘Brigitta Blue’, V. virgatum FLW84-52, V. amoenum CVAC1155.
Experimental design and analysis.
Berries of available species and genotypes were isolated with flies in 2 – 3 separate plastic vials arranged in a fully randomized design according to a no-choice bioassay, i.e., only berries of one genotype were placed in each individual vial. Berries were dissected individually and were treated as experimental replicates. Effects of genotype, ploidy level, berry weight and brix on fly reproductive parameters (i.e., eggs, larval density, female brood output, reproductive rate, and adult mortality) were analysed with correlation analysis (Proc CORR in SAS 9.4) and 1-way or 2-way multivariate analysis of covariance (ANCOVA) with berry weight and brix serving as covariates (Proc GLM in SAS 9.4). Resistance indices were compared among species using one-way analysis of variance (1-way ANOVA, Proc GLM in SAS 9.4) with mean separations accomplished with Tukey’s HSD tests at an α = 0.05. Likewise, the effect of adult mortality on oviposition rate was assessed using 1-way ANOVA after arcsine-square-root transforming adult mortality rate. Cluster analysis (PROC ACECLUS in SAS 9.4) followed by phylogenetic analysis using the first two canonical variables (PROC TREE in SAS 9.4) grouped species and associated genotypes by clutch size (eggs per berry), larvae per berry, female reproductive output, and resistance index. We adjusted female reproductive responses for variation in berry size (i.e., weight). Non-linear relationships between adult mortality and egg count, resistance index and adult mortality, ploidy and brood output, ploidy and fruit weight, ploidy and antibiosis, fruit weight and reproductive rate, fruit weight and brood output, and fruit weight versus brix were modelled using PROC NLIN in SAS 9.4.