Behavioral evidence for contextual olfactory-mediated avoidance of the ubiquitous phytopathogen Botrytis cinerea by Drosophila suzukii
Abstract
Herbivorous insects may benefit from avoiding the smell produced by phytopathogens infecting plant host tissue if the infected tissue reduces insect fitness. However, in many cases the same species of phytopathogen can also infect host plant tissues that do not directly affect herbivore fitness. Thus, insects may benefit from differentiating between pathogen odors emanating from food and nonfood tissues. This is based on the hypothesis that unnecessarily staying attentive to pathogen odor from nonfood tissue may incur opportunity costs associated with not responding to other important survival functions. In this study adults of Drosophila suzukii Matsumura, an invasive larval frugivore, showed reduced attraction to the odor of raspberry fruit, a food tissue, when infected with Botrytis cinerea Pers., a ubiquitous phytopathogen, in favor of odors of uninfected raspberry fruit. Moreover, D. suzukii oviposited fewer eggs on infected raspberry fruit relative to uninfected raspberry fruit. Larval survival and adult size after eclosion were significantly reduced when reared on B. cinerea-infected raspberry relative to uninfected fruit. Interestingly, when the behavioral choice experiment was repeated using Botrytis-infected vs. -uninfected strawberry leaves, a nonfood tissue, in combination with fresh raspberry fruit, odor from B. cinerea-infected leaves did not reduce D. suzukii attraction to raspberries relative to raspberries with uninfected leaves. These behavioral results illustrate the important role context can play in odor-mediated interactions between insects, plants and microbes. We discuss implications of our findings for developing a repellent that can be useful for the management of D. suzukii.
Introduction
Recent interest in interkingdom chemical communication highlights the role of microbial volatile semiochemicals in mediating insect host choices (Davis et al., 2013). The majority of such studies have focused on elucidating the role of microbial volatiles in insect attraction to host plants. For example, it has been suggested that the attraction to fruit in Drosophila melanogaster Meigen is mediated by volatiles produced by the yeasts present on the fruit phyllosphere (Becher et al., 2012). However, specific host choice involves both attraction to a suitable host and avoidance of unsuitable hosts. For example, it is well known that in Rhagoletis pomonella (Walsh), both attraction and avoidance responses to certain odors from natal and nonnatal host fruit mediate host choice behaviors of different R. pomonella host races (Linn et al., 2003; Linn et al., 2005; Feder & Forbes, 2007; Linn et al., 2012). While the role of microbial semiochemicals in the attraction of phytophagous insects has gained recognition (Davis et al., 2013), less is known about their ecological and behavioral influences in mediating insect avoidance of a non- or unsuitable hosts.
Avoidance of potentially harmful odor has been demonstrated in many animal systems (Meisel & Kim, 2014; Li & Liberles, 2015). In the model insect D. melanogaster, for example, one such avoidance behavior quantified in the form of reduced attraction was shown to be regulated by a special olfactory circuit dedicated to process an earthy smelling volatile terpenoid, geosmin, produced by a number of different bacteria, fungi, and cyanobacteria (Stensmyr et al., 2012). Such avoidance behavior is presumably adaptive due to negative fitness consequences exerted by geosmin and the associated microbes. For example, Penicillium spp., one of the geosmin producing fungi, has been shown to produce insecticidal and insect antijuvenile hormone metabolites (Castillo et al., 1999). In another example, Botrytis cinerea Pers., a ubiquitous necrotrophic plant pathogen, has been shown to induce olfactory avoidance behavior from adult Lobesia botrana (Denis & Schiffermuller) to its host fruit, grape. This avoidance behavior is associated with reduced larval performance in the infected fruit (Tasin et al., 2012).
Microbes potentially harmful to insects, such as Penicillium and B. cinerea discussed above, are ubiquitous in nature, colonizing a wide range of habitats from soil to plant, and emit volatiles where they grow. Hence, these microbes can colonize both plant tissue that is a resource for the insect (e.g., food tissue) and plant tissue that is not directly used by the insect (e.g., nonfood tissue). This suggests that insects may benefit from having the capacity to differentiate between odors from harmful microbes that are emanating from tissues used or not used by the insect as a food source. This prediction is based on the argument that remaining attentive to a false or irrelevant signal will have negative trade-offs in other important survival functions due to the limited neural capacity of insects (Bernays, 2001). As the volatile profile of microbes can be significantly influenced by the substrates they grow on (Ferreira et al., 2000; De Lucca et al., 2012), we hypothesize that a phytophagous insect will be able to differentiate between volatile metabolites produced by a potentially harmful microbe emanating from host tissue used by the insect as adult food or as a reproductive host compared to volatile metabolites from the same microbe emitting from plant tissue not used as a food resource.
Spotted wing drosophila (Drosophila suzukii Matsumura) is currently a serious invasive pest of raspberries, blueberries, strawberries, and cherries in the United States and Europe. In contrast to D. melanogaster that mainly colonize overripe or rotten fruit, D. suzukii evolved a behavioral and morphological innovation (enlarged, serrated ovipositor) that enables gravid females to oviposit on fresh fruit in addition to overripe fruit (Lee et al., 2011; Atallah et al., 2014). Botrytis cinerea is an important pathogen of brambles and other fruits and can infect flowers, ripe fruit, and vegetative tissue causing a disease referred to as gray mold (Elad et al., 2004). We hypothesized that fruit infected with B. cinerea would compete with and negatively affect D. suzukii larval performance as shown in other insects such as L. botrana (Tasin et al., 2012). Moreover, based on the arguments developed above, we further hypothesized that D. suzukii adults are able to behaviorally avoid odors from B. cinerea-infected fruit but not respond to odors from B. cinerea-infected leaf tissue.
Here we examine (i) whether adult D. suzukii show reduced attraction to an attractive odor if combined with volatiles produced from B. cinerea, (ii) whether the reduced attraction is ultimately correlated with negative effects of B. cinerea on D. suzukii performance, and lastly (iii) whether the reduced attraction is context dependent; that is, whether D. suzukii’s response to volatiles from B. cinerea-infected tissue is dependent on whether the fungus is infecting food tissue (e.g., fruit) vs. nonfood tissue (e.g., leaf). To answer these questions, we manipulated B. cinerea on raspberry fruit, a preferred reproductive host of D. suzukii (Lee et al., 2011) and a host of B. cinerea (Farr et al., 1989), and also on strawberry leaves, a host of B. cinerea (Braun & Sutton, 1988). Drosophila suzukii adults have been shown to be attracted to strawberry leaves (Keesey et al., 2015). Specifically, we compared olfactory-based preference, oviposition preference, and larval performance of D. suzukii to raspberry fruit with and without B. cinerea, and tested olfactory preference of D. suzukii to a combination of raspberry fruit + strawberry leaf vs. a combination of raspberry fruit + strawberry leaf infected with B. cinerea.
Materials and methods
Insect colony
Drosophila suzukii used in behavioral, oviposition, and larval performance experiments were originally collected from the wild in central New York, USA in 2014 and subsequently reared on a standard cornmeal diet (1 L distilled water, 40 g sucrose, 25 g cornmeal [Quaker Oats Co., Chicago, IL, USA], 9 g agar [No. 7060, Bioserve, Flemington, NJ, USA], 14 g torula yeast [No. 1720, Bioserve], 3 mL glacial acetic acid [Amresco, Solon, OH, USA], 0.6 g methyl paraben [No 7685, Bioserve], and 6.7 mL ethanol), in plastic bottles with foam stoppers (Applied Scientific, Carson, CA, USA) in a climate controlled chamber at 25 ± 1 °C, 60% ± 5 % RH, 16 : 8 L : D (Wallingford et al., 2016). Newly eclosed flies were collected from bottles and moved to fresh media every 24 h and allowed to mature for 4–7 d before using them in adult assays. To obtain eggs for larval performance assays, mature adults (approx. 250 mixed gender) were isolated in 300 mL plastic rearing bottles with Carolina blue media (11.6 g in 44 mL distilled water) and allowed to oviposit for 18 h after which the adults were removed. Eggs were freed from media using a previously reported procedure (Schou, 2013) with modifications. Briefly, 50 mL of a sucrose solution (29 g sucrose/100 mL water) was added to each bottle with eggs, agitated for 10 s and then decanted into a 100 mL beaker where the solution was agitated for an additional minute followed by adding an additional 25 mL of the sucrose solution. After settling, most eggs floated to the top of the solution in the beaker where they were moved to moist filter paper within a glass funnel using a pipet. A fine brush, rinsed with 95% ethanol, was used to move eggs to fruit as described below.
Plants
Organically produced ripe red raspberries used in behavioral, oviposition, and performance assays were purchased from a local grocery as needed. Strawberry leaves used for behavioral assays were obtained from potted strawberry plants grown in the greenhouse at Cornell AgriTech in Geneva, NY, USA.
Botrytis cinerea
Isolates of Botrytis cinerea used in this study were obtained from organic strawberries purchased in a local grocery store. Briefly, strawberries were placed in individual moist chambers for 4 d at 22 °C under a 12 : 12 L : D. Conidia characteristic of B. cinerea were scraped from the fruit and suspended in 1 mL sterile distilled water. To obtain single-conidial isolates, 100 μL of each resulting suspension was evenly distributed onto potato dextrose agar (PDA; Difco Laboratories, Detroit, MI) and incubated for 24 h. Following incubation (22 °C, 12 : 12 L : D) germinated conidia were selected, transferred to PDA and incubated at the conditions described above for 7 d. Conidia from each single-conidial colony were suspended in sterile distilled water and stored at −20 °C until use.
For fruit and leaf inoculation, a single-conidia isolate of B. cinerea (strain 3RGA33B) was removed from storage and 75 μL of the thawed conidial suspension (∼103 to 104 conidia/mL) was evenly distributed on PDA. Following 1 week of incubation, conidia were scraped using a sterile probe and suspended in 1 mL of sterile distilled water. Density of spores was not quantified and probably varied across inoculations.
All inoculations of fruit with B. cinerea were carried out in a laminar flow hood. The surface of ripe raspberries was first sterilized with 35% ethanol solution followed by a triple rinse with sterile distilled water. For each raspberry, a small wound was created with a sterile scalpel and was subsequently inoculated with either 25 μL of either the conidial suspension or sterile distilled water. All inoculated raspberries were immediately placed individually in 30 mL plastic shot glasses with lid that were previously rinsed with 70% ethanol and then put in a growth chamber (25 °C, 50%–60% RH, 16 : 8 L : D) until used for bioassays (see below).
Leaf inoculations with B. cinerea were carried out following previously published procedures (Rizvi et al., 2015). Green, healthy, mature strawberry leaflets (unknown age, 60 mm by 50 mm dimensions) were collected from potted strawberry plants, surface sterilized with 1% NaClO solution for 5 min and rinsed with sterile water three times. The surface of each nonwounded leaf was inoculated with either 25 μL of the conidial suspension used for fruit inoculations or sterile distilled water. Immediately following inoculation, individual leaves were placed in zip lock plastic bags (350 mm × 400 mm) with damp sterile filter paper and incubated at 22 °C and 12 : 12 L : D conditions. After 10 d of incubation, fungal lesions were observed on leaves inoculated with the conidial suspension only. Visual confirmation of B. cinerea infection was accomplished by excising small pieces of symptomatic tissue from inoculated leaves and culturing on PDA.
Effect of B. cinerea on D. suzukii emergence and growth
Raspberries were inoculated (Botrytis treatment) or sham-inoculated (water control) 24 h prior to the experiment. On the day of the experiment, a small wound was made on the side of each raspberry using a sterilized scalpel approximately 5 mm away from wound made for Botrytis inoculation or water control. Five eggs were moved into the wound site per fruit using a fine brush sterilized with 95% ethanol. Three raspberries (total of 15 eggs) were placed in a 120 mL plastic rearing cup with screened top and bottom for air circulation and drainage of liquid, respectively (n = 10). Cups were placed in a growth chamber set at 25 ± 1 °C, 60% ± 5% RH, 16 : 8 L : D. After 4 d, cups were checked daily and eclosed adults were collected and placed in 95% ethanol until measurement by replicate. Individual flies were measured for wingspan and thorax length using a dissecting scope equipped with an ocular micrometer and weighed after dried. Wingspan was the distance between the anterior portion of the wing base and the distal wing margin at the L3 vein (Wallingford & Loeb, 2016). Thorax length was the distance between the anterior margin of the thorax (propleuron) and the posterior tip of the scutellum (Wallingford & Loeb, 2016).
Effect of B. cinerea on D. suzukii oviposition
For this experiment, raspberries inoculated with B. cinerea 48 h prior to the test were used to minimize potential physical interference of the fungal mycelia on D. suzukii oviposition behavior. The two-choice experiment was conducted as described below. Two raspberries (one water-inoculated and one B. cinerea-infected) were placed on a petri dish (100 mm diameter) in a cage. Ten 6-d-old mated D. suzukii females were added to each cage and allowed to oviposit for 5 h. The test was set up at 11 am and eggs were counted using a dissecting microscope at 4 pm with a total of 16 replicates.
Laboratory two-choice bioassays: general protocols
We conducted two experiments using a two-choice assay as described in Cha et al. (2012) and as similarly shown in Fig. S1C. Briefly, the experiment was conducted in plastic insect cages (600 mm W × 600 mm L × 600 mm H; BugDorm-2120 Insect tent; shop.bugdorm.com). Within each arena were two gated traps (treatment and control) positioned at the center of arena, 100 mm apart from each other. We used a 250 mL (for the first two-choice tests) or 600 mL (for the second two-choice test) glass beaker covered with aluminum foil with a cut centrifuge vial (7 mm diameter) inserted in the foil for D. suzukii entry, as a trap. Each trap contained 30–60 mL soapy water (0.0125% odor free dishwashing soap, seventh generation, www.seventhgeneration.com) as a drowning solution. On average 100 flies (roughly 1 : 1 male : female ratio, 4–7 d old) were released in each arena. A cotton ball (30 mm diameter) soaked with distilled water was placed in the center of each arena to provide water for adult flies. Choice assays started at 1 pm each day (21.5 ± 0.9 °C, 60% ± 5% RH, 16 : 8 L : D). The number of male and female flies inside treatment and control traps was counted after 20 h.
Choice test 1: D. suzukii behavioral response to odor of B. cinerea grown on raspberry
Each cage contained a control trap and a treatment trap (n = 20). The control trap contained two water-inoculated raspberry fruit enclosed in two 30 mL deli-cups (one fruit/cup) with a screen on top to prevent direct contact of flies with fruit. The treatment trap contained one water-inoculated raspberry fruit and one B. cinerea-infected raspberry fruit with each fruit enclosed in a separate 30 mL deli-cup with a screen on top. Raspberries were inoculated with B. cinerea or sterile distilled water and tested 3 d after inoculation, when fruit surface tissue degradation is visually noticeable on Botrytis inoculated fruit but without visible mycelia growth. Mycelia growth was fully visible on the infected raspberries by the end of the experiment.
Choice test 2: D. suzukii behavioral response to odor of raspberry in combination with odor of B. cinerea grown on strawberry leaf
Each cage contained a control trap and a treatment trap (n = 10). The control trap contained a fresh raspberry fruit enclosed in a 30 mL deli-cup with a screen on top and a water-inoculated strawberry leaflet enclosed in a 30 mL deli-cup with a screen on top. Treatment traps contained a fresh raspberry fruit enclosed in a 30 mL deli-cup with a screen on top and a strawberry leaflet infected with B. cinerea enclosed in a separate 30 mL deli-cup with a screen on top. Strawberry leaflets were inoculated with B. cinerea or water and tested 11 (n = 4) and 12 (n = 6) days after inoculation. Fungal mycelia were visible on the leaf surface at the time of the experiment for leaflets inoculated with B. cinerea.
Statistical analysis
Adult emergence data were analyzed using a completely randomized design with the Botrytis treatment as a fixed factor using Proc Mixed in SAS (SAS Institute, 2009). Performance (e.g., thorax length, wingspan, and body weight) data were analyzed using a completely randomized design with the Botrytis treatment, sex and the interaction as fixed factors using Proc Mixed in SAS (SAS Institute, 2009). Oviposition data and trap catch data were analyzed using generalized linear models in a randomized block design with block as a random factor and Botrytis treatment, sex, and the interaction as fixed factors using Poisson distribution with log link function and maximum likelihood estimation (Proc Glimmix, SAS Institute, 2009). The means were compared using the Tukey–Kramer test (SAS Institute, 2009).
Results
Effect of B. cinerea on D. suzukii emergence and growth
Adult D. suzukii emergence from raspberry fruit was significantly reduced by B. cinerea infection on fruit (Fig. 1). Among the 15 eggs transferred to 3 raspberry fruit, 14.4% fewer adults emerged per fruit from fruit infected with B. cinerea compared to sham-inoculated fruit (12.4 ± 0.60 and 14.4 ± 0.22, respectively; F1,18 = 9.78, P = 0.0058). Adult growth characteristics (e.g., thorax length, wingspan, and body weight) of emerged adults were also negatively influenced by B. cinerea, regardless of the sex (main effect of B. cinerea inoculation for all three measurements: P < 0.0001), and females were significantly larger than males regardless of B. cinerea inoculation (main effect of sex for all three measurements: P < 0.0001) (Fig. 2). There were no significant interactions between B. cinerea and sex for all three growth measurements (for all three P > 0.28). Specifically, on average, female flies that emerged from B. cinerea-infected fruit were 6.4% smaller in thorax length, 4% shorter in wingspan, and 9.6% lighter in dry weight compared to female flies emerged from sham-inoculated fruit. Similarly, male flies emerged from B. cinerea-infected fruit were 5.7% smaller in thorax length, 4.1% shorter in wingspan, and 12.5% lighter in dry weight compared to male flies emerged from sham-inoculated fruit.
Effect of B. cinerea on D. suzukii oviposition
In a choice test, female D. suzukii oviposited significantly fewer eggs on raspberry fruit infected with B. cinerea compared to sham-inoculated raspberry fruit (F1,15 = 23.45, P = 0.0002, Fig. 3). At the tested stage (48 h after inoculation), mycelial growth on the outside of the infected fruit was still not visible. Thus, the observed reduction in oviposition was not likely due to the mycelial growth that could physically interfere with oviposition at a later postinoculation stage.
Effect of B. cinerea odor on adult D. suzukii attraction
Attraction of D. suzukii to the odor of raspberry fruit was significantly reduced when fruit were infected with B. cinerea (Fig. A). Traps baited with a fruit sham-inoculated with water caught significantly greater numbers of D. suzukii than traps baited with B. cinerea-infected fruit (F1,57 = 255.99, P < 0.0001) and the response was similar for both males and females (F1,57 = 0.02, P = 0.8939). In contrast, we did not observe reduced attraction to B. cinerea odor by adult D. suzukii when we presented odors of B. cinerea grown on strawberry leaves in combination with fresh raspberry fruit (Fig. B). The numbers of flies captured were similar in traps containing strawberry leaflets with sham-inoculation or B. cinerea-infected leaves, where both traps also contained a sham-inoculated raspberry fruit (F1,27 = 0.97, P = 0.3341). Although there was a significant overall effect of sex (male vs. female: 10.9 ± 1.3 vs. 13.3 ± 1.3; F1,27 = 5.01, P = 0.0337) on trap catches, mean comparisons revealed no significant differences among different treatment combinations (Fig. B).
Discussion
In this study behavioral choice experiments indicate that D. suzukii shows reduced attraction to reproductive hosts and larval food hosts infected by B. cinerea via olfactory-mediated avoidance behavior. Three lines of evidence support the conclusion that the reduced attraction represents avoidance behavior in terms of outcomes. First, other studies have shown that B. cinerea reduces insect fitness and the insect avoids B. cinerea odor (Tasin et al., 2012). Second, numerous studies have concluded olfactory-mediated avoidance behavior in insects based on similar bioassay approaches measuring reduced attraction (Stensmyr et al., 2012; Pham & Ray, 2015; Du et al., 2016; Renkema et al., 2016). Third, we overcompensated for the amount of fruit volatiles (i.e., attractant) in the Botrytis treatment (more detail below, Fig. S1) and fewer D. suzukii were captured in the Botrytis treatment that had at least the same amount, if not more, of fruit volatiles than the control trap baited with raspberry infused agar. Interestingly, in our study, the reduced attraction depended on where B. cinerea grows, suggesting that D. suzukii may be able to distinguish B. cinerea growing on strawberry leaves (a potentially false warning) from B. cinerea growing on raspberry fruit (a real fitness threat). While the chemical and neurological mechanisms underlying the reduced response is still unclear, this type of context dependency in behavioral response is an expected finding for D. suzukii based on fitness arguments. Although B. cinerea on fruit negatively impacts D. suzukii fitness and thus the observed reduced attraction to infected fruit is likely adaptive, reducing attraction to uninfected fruit when B. cinerea growing on leaf tissue is present in the immediate environment may be maladaptive. Displaying reduced attraction more often than necessary based on a false olfactory warning may incur opportunistic costs that further diminish the fitness outcome of D. suzukii (Bernays, 2001).
An alternative explanation for the reduced attraction to treatments including raspberries infected with Botrytis is that the fungus could have reduced the quantity of attractive odors being produced by the fruit and therefore the differential capture was simply due to lower levels of attraction. However, our preliminary experiment using pieces of raspberry infused PDA medium (i.e., raspberry agar, RA) infected with and without B. cinerea showed lower level of attraction even when the Botrytis treatment had at least equivalent, if not more, quantity of fruit odor as the control (Fig. S1). Specifically, the control treatment contained a piece of RA, while the Botrytis treatment contained a piece of RA plus a piece of RA infected with B. cinerea (Fig. S1C), suggesting that the reduced captures in the Botrytis treatment trap was due to volatiles emitted from B. cinerea-infected fruit agar rather than lower quantity of attractant (Fig. S1D), although it is also possible that some antennally active Botrytis volatiles could have rendered otherwise attractive raspberry volatiles unrecognizable to D. suzukii.
Since we have not identified the actual behaviorally active chemicals released from Botrytis-infected fruit yet, currently we can only speculate on the potential chemical mechanisms underlying differential responses of D. suzukii to volatiles released from Botrytis-infected raspberry fruit and strawberry leaf. It may be possible that raspberry fruit colonized by B. cinerea produces a repellent chemical to D. suzukii but the repellent is not produced from strawberry leaf colonized by B. cinerea. It is also possible that B. cinerea on both raspberry fruit and strawberry leaves produce the same repellent chemical but D. suzukii is adapted to recognize some additional volatile cues from strawberry leaves that override the avoidance response or to simply filter out the irrelevant signals (Von der Emde & Warrant, 2016). Another explanation would be that the B. cinerea infection masks the signal from food tissues, thereby decreasing attraction without explicit avoidance behavior. Similarly, it is also possible that the odor blend from the infected fruits is not recognized as a host signal because of the inclusion of fungal volatiles in the blend (Bruce & Pickett, 2011). Also, as is the case with many attractant cues that involve multiple volatile components rather than a single attractant compound, specific repellent cues may involve a blend of chemicals. The volatile profiles of B. cinerea-infected raspberry fruit or strawberry leaf may differ in composition, ratio, and/or concentration (Bruce et al., 2005; Cha et al., 2011). We also recognize the possibility that the amount of volatiles produced by Botrytis-infected fruit and leaf may be different (e.g., lower amount of volatiles from Botrytis-infected leaf than fruit) and the amount of potential repellent from Botrytis-infected leaf is simply under a behavioral threshold in our experiment and not recognized by D. suzukii. Ongoing work on the identification of the Botrytis based repellent will help to elucidate the actual chemical mechanism.
Although the magnitude of the negative impact of B. cinerea on D. suzukii performance was small, the effect was consistently significant. This suggests that the reduced attraction to odor from fruit infected with B. cinerea is correlated with reductions in D. suzukii oviposition and performance. This type of positive association between preference and performance (Craig & Itami, 2008) was also observed for European grape berry moth and grape infected with B. cinerea in which adults avoid infected grape and the infected grape supports reduced larval performance (Tasin et al., 2012). Still, in our lab assays, some individual flies did not show reduced attraction for odors from Botrytis in raspberry fruit and oviposited on Botrytis-infected raspberry, raising the interesting question on the potential role of genetic variation or physiological state of D. suzukii, such as microfauna, on the response to B. cinerea odor.
In conclusion, although the exact chemical and neurophysiological mechanisms underlying the reduced attraction of D. suzukii adults to B. cinerea odor still needs elucidation and is underway, our results provide behavioral evidence consistent with context dependence in the olfactory-mediated avoidance of harmful microbes by D. suzukii. In practical terms, identifying the chemicals responsible for the avoidance behavior is an important next step that could lead to the development of new repellents as part of a more sustainable approach to the management of D. suzukii.
Acknowledgments
We thank Charlie Linn and four anonymous reviewers for insightful comments. This research was supported in part by funding from Cornell University's New York State Agricultural Experiment Station federal formula funds, project 2015-16-180, New York State Agriculture and Markets (C0011GG) and National Institute of Food and Agriculture (2016-0228-08).
Disclosure
All authors are without conflicts of interest, including specific financial interests and relationships and affiliations relevant to the subject of this manuscript.
