Source: UNIVERSITY OF CALIFORNIA, RIVERSIDE submitted to NRP
IDENTIFICATION OF ODOR-BASED SPATIAL REPELLENTS FOR CONTROL OF AGRICULTURAL PESTS
Sponsoring Institution
National Institute of Food and Agriculture
Project Status
COMPLETE
Funding Source
Reporting Frequency
Annual
Accession No.
1018395
Grant No.
(N/A)
Cumulative Award Amt.
(N/A)
Proposal No.
(N/A)
Multistate No.
(N/A)
Project Start Date
Dec 7, 2018
Project End Date
Sep 30, 2022
Grant Year
(N/A)
Program Code
[(N/A)]- (N/A)
Recipient Organization
UNIVERSITY OF CALIFORNIA, RIVERSIDE
(N/A)
RIVERSIDE,CA 92521
Performing Department
Molecular, Cell and Systems Biology
Non Technical Summary
Current methods of controlling invasive pest insects such as the spotted wing drosophila and navel orangeworm largely depend on pesticides. Using a novel computational method we have developed, we will evaluate >350,000 natural compounds to identify ones that can act as strong repellents to these insects. We propose to create odor-based insect repellents that act over a wide region for affordable, safe, ways to reduce contact between agricultural pests and their plant hosts.
Animal Health Component
70%
Research Effort Categories
Basic
30%
Applied
70%
Developmental
0%
Classification

Knowledge Area (KA)Subject of Investigation (SOI)Field of Science (FOS)Percent
21109991130100%
Goals / Objectives
The goal of this proposal is to identify non-toxic, affordable and natural odorants that can repel a broad spectrum of agricultural pests from plants. Most insects rely on the sense of smell for a number of important behaviors like finding food and avoiding predators. Pest insects, whichdestroy a large portion of the global agricultural produce by feeding on them, identify their host plants using various cues, including volatile odorants that these plants emit. Some plant feeding insects can also transmit deadly diseases like the Citrus Greening Disease. Yet little is known about the olfactory system in these pest species, particularly about exploiting them for behavior control.There are few repellents that are effective against biting arthropods such as diethyl toluamide (DEET), but none that are useful for plant pests. There is a need to identify other efficient compounds that may be safe to use. The current methodsof control of plant pests relies largely on insecticides. Heavy use of these chemicals can have adverse effects on the environment and on human health. Additionally, it can also lead to insects that are resistant to these insecticides. Other methods of control that are used less frequently include traps, repellents and mating disruption using pheromones.Insects use their olfactory organs and this is important for a number of different important behaviors including finding their host plants(Clark and Ray, 2016; van der Goes van Naters and Carlson, 2006). These volatile odors are detected by olfactory sensory neurons that are housed in hair-like structures called sensilla. Odorant molecules are recognized on the surface of the neurons by 7-transmembrane receptors that are encoded by members of two types of large gene families, the Odor receptor (Or) gene family and the Ionotropic receptor (IR) gene family. The activation of olfactory neurons by odorants leads to an increase in the frequency of action potentials generated by that neuron which can be measure using single-unit electrophysiology(Hallem et al., 2006).Identification of a highly conserved repellent receptor in insects:There are major impediments to finding powerful insect repellents. Although DEET has remained the top repellent for human use against biting insects, nothing comparable is used widely for agricultural insects. Identifying improved DEET alternatives, particularly ones that work against agricultural pests, has been daunting because the receptors causing olfactory repellency are unknown, and new chemistries require exorbitant costs to determine human safety before use. We have identified the mode of action of DEET in the olfactory system (unpublished) and it acts broadly as an agonist across several classes of olfactory neurons. This broad mode of actions makes it difficult to use for screening of new repellents using physiological assays.Another major limitation in developing improved repellents is the high costs (>$10-30M) anticipated for identification and determination of safety for use. The repellents we are proposing to test will overcome this limitation in several ways. They will be natural compounds, ideally commonly used in human food and cosmetics, excellent safety profiles with significantly better toxicity values than pesticides, preferably listed Generally Regarded as Safe (GRAS) by Food and Extract Manufacturers Association (FEMA).We have developeda high-throughput repellent ligand screening system for insects using chemical informatics:We used a structure-matching chemical informatics method we had created earlier to screen in silico a large chemical library consisting of >440,000 volatile-like chemicals including ~12,000 chemicals identified as either originating from plants, insects, or vertebrate species. These also include several compounds already approved for human use as fragrances, cosmetics or flavors(Boyle et al., 2016a; Boyle et al., 2013; Boyle et al., 2016b; Tauxe et al., 2013). We identified the top 150 predicted repellent compounds and arranged them by predicted logP and vapor pressure values to provide a high-priority list of candidates for behavioral testing. Nearly 60 of the predicted compounds were tested for repellency usingDrosophila melanogasterandAedes aegyptiand we observed a success rate of nearly 70%.A more limited number of predicted chemicals were tested more broadly across multiple insect species and showed a strong repellency toDrosophila melanogasterin a 2-choice assay as well as toDrosophila suzukii. For example, we released D. suzukii in large glass chambers and offered them a choice between 2 sets of blueberries, one set coated with test compound (BA at 10%) and the other coated with solvent. After 2 weeks the blueberries were isolated and adults emerging counted. The blueberries were also dissected, and eggs and larva counted. The treated grapes had a dramatic reduction in number ofD. suzukiiat all stages suggesting that the repellency was extremely effective(Pham and Ray, 2015). Argentine ants and bed bugs also showed strong repellency to these compounds.However, these insect repellents are based on low volatility compounds (DEET, picaridin, PMD) and therefore act only at a close-range of < 1 cm necessitating careful spraying on the whole plant for complete coverage. A repellent that showed repellency at a greater distance would enable more effective protection and not require such extensive spraying. If we can discover natural odorants that shows spatial ranges of repellency (>5cm and up to 1 m), and is safe from toxicologyperspective, then we could use it in ways that do not require high application rates on our crops and food.Thehighest priority for this proposal is to find actives with improvements in greater repellent spatial distance (> 1 m) and stronger repellency. We propose toenhance the screening method by using a more advanced computational pipeline we implemented only recently.Importantly, we will identify repellents have predicted vapor pressures >10-100 times that of the initial round of repellents and therefore likely to provide spatial protection across a distance. Care will be taken to empirically balance the spatial range with the duration of protection such that the compound does not dissipate in less than 24hrs time. The dose of volatile compounds can be controlled also by delivering them from canisters and sprayers with larger quantities. These chemicals will also have desirable smells that can be formulated into floral, fruity, herbal smelling formulations that do not take away from the food quality.We expect that the pleasurable scent imparted in the form of a perfume to use on the body or around the hut will significantly improve adoption and use.The vapor pressures of the test compounds are predicted to have a large range, raising additional possibility for development into different delivery formats besides a clip-on dispenser.Aim 1.Computationally screen 360,000 natural chemicals for predicted repellents and prioritize compounds with high vapor pressure.Aim 2.Behavioral testing of predicted repellentsonDrosophila melanogasterfor prioritization of spatial repellents.Aim 3.Testing of compounds with ideal usability properties for use on crops.Aim 4.Spatial repellency testing of repellentsonDrosophila suzukii.Aim 5.Repellent testing forAmyelois transitella, the navel orangeworm.
Project Methods
Aim 1.Computationally screen 360,000 natural chemicals for predicted repellents and prioritize compounds with high vapor pressure.We will screen a proprietary library of 350K natural compounds compiled from a variety of metabolite databases including plant, human, yeast, bovine amongst others.Significantimprovementswill be achieved in 3 ways: ~doubling the number ofphysicochemical descriptors used (from 3200 to 6000), ~ doubling the trainingset of knownrepellents (from 60 actives to 130 actives with more diversestructures), and using a 30 times larger natural compound library to screen (from12,000to 360,000).Theincrease in physicochemical descriptors is due to an update for the DragonSoftware package, which added several 2-D and 3-Ddescriptors in its newerrelease.Increase in "trainingset" size and diversity by adding thenewly identified 70 repellents noted in the previous work section.The top 500 hits from the screen will be evaluated for Vapor pressure (known or computed) and for human safety (known or computed using EPI Suite software).The top 20 compounds will be selected with Vapor pressures >10-100 times that of existing anthranilate repellents and predicted safety values better than LD50s>3500mg/kg. The compounds will also be separated into bins according to their known (or predicted) smells: fruity, citrus, vegetable, earthy, woody and other pleasant. We can predict the smell of any chemical from structure alone in our laboratory using a machine learning software we developed. The top 50 compounds will be identified by prioritizing the hits according to 4 criteria: predicted repellency, vapor pressure, low toxicity, and suitable smell for use in agriculture. Additionally their availability in pure form from an existing source will be evaluated.Aim 2. Behavioral testing of predicted repellentsonDrosophila melanogasterfor prioritization of spatial repellents.We propose to validate the repellency properties of 50 of the selected candidates againstD. melanogasterwhich is easy to test in the laboratory. Our previous experience has shown a strong correlation between repellency seen withD. melanogasterand other insect species (mosquitoes, ants, ACP).We will carry out two behavioral assays: (1) medium-range of ~6 cm assay for reduction in vinegar-baited trap entry, and (2) long-range reduction ~ 100 cm in in vinegar-baited trap entry. We will evaluate which repellent chemicals can provide the maximum efficacy of protection from a distance to the trap entry point.Chemicals will be acquired at the highest purity available from a commercial inventory source such as eMolecules which includes chemicals from hundreds of suppliers. The vinegar baited traps have been used in the past and include a small funnel fashioned out of a cut 1ml pipette tip inserted into an upside down 2 ml microcentrifuge tube. The fly can enter the trap but is unable to leave. A small amount of apple cider vinegar (0.1ml) is added to the base inside the microcentrifuge tube cap as a strong lure and placed in a large glass arena. The test odorant or solvent control will be placed on a piece of filter paper (10% solution in ethanol or water, 0.1 ml) on a disposable foil piece placed 5 cm in front of the trap face. Approximately 20 flies (10 males and 10 females) will be released in the chamber and allowed to behave for 24 hours in a light/dark cycle. The number of flies entering the trap will be noted and the percentage repellency compared to control calculated for each of the 50 compounds.The compounds showing >90% repellency will be further tested at the longer distance of 100cm. A similar trap will be deployed but inside a large indoor sealed greenhouse. The trap will be surrounded by 4 odorant filter papers at the 100cm distance in the 4 directions. As before the released flies will be allowed to behave for 24 hours and the numbers entering the traps counted. The percentage repellency values will be calculated and the top 15 candidates identified.Aim 3.Testing of compounds with ideal usability properties for use on crops.In order to assess whether the top 15 candidate compounds have a harsh effect on the plants themselves a few tests will be performed. First, a phytotoxicity test will be performed usingArabidopsis thalianaseedlings as a model. While each compound will eventually need to be tested on the target plant species, these initial tests will provide useful information to triage the compounds showing high phytotoxicity. Each repellent compound will be applied directly onto the leaves as a 10%, 3% and 1% concentration along with solvent as a control. The plants will be observed for 1 week and photographs taken to compare the yellowing of leaves, leaf death, and height of seedling. The process will be repeated at 30 Hrs and after 1 week.The compounds identified with low phytotoxicity will be tested for photostability when applied on a filter paper and left outdoors exposed to the environment. The compounds will be tested for concentration changes after 1 day, 2 days, 1 week, and 2 weeks using a gas chromatogram (GC) linked Mass Spec (MS).Aim 4.Behavioral testing of predicted repellentson D. suzukii.The 6 compounds will be tested as repellents in an assay where we evaluate the reduction in oviposition on blueberries. We will evaluate which repellent chemicals can provide the maximum efficacy of protection againstD. suzukii, at the furthest distance, at the lowest concentration.In order to assess the sensitivity to the repellency-mediated by the spatial repellents, dose response assays will be performed. Each repellent will be tested at four different concentrations (10%, 5%, 1%, 0.5%) in two different assays.In the first test the 6 best candidates will be tested in the glass arena using a two-choice blueberry test. One set of blueberries will be sprayed lightly with the test compound and the other with the solvent (water or ethanol). Approximately 20 mated femaleD. suzukiiwill be released. After 10 days the blueberries will be dissected under a microscope and the numbers and stages of larvae and eggs counted.In the second test each of the 6 compounds will be tested in a formulation called SPLAT (ISCA technologies). The sticky dollops of SPLAT mixed with the repellent will be applied at a distance of 50 cm from the location of the blueberries in a sealed indoor arena of 9 ft X 10 ft x8 ft. Approximately 20 mated femaleD. suzukiiwill be released. After 10 days the blueberries will be dissected under a microscope and the numbers and stages of larvae and eggs counted.Using these behavioral tests we expect to identify the most effective repellent compound that is affordable, safe and pleasant smelling. The results from these tests will inform us of the possibility to perform large-scale field trials for fruit protection in the future.Aim 5.Repellent testing forAmyelois transitella, the navel orangeworm.In our previous work we found that that most repellents act across multiple species of insects and even on more distantly related arthropods. We will test the top 3 repellents identified in Aim 4 as oviposition deterrents, at the lowest effective concentrations, on the navel orangeworm adults. Repellent odorants will be tested in a cage assay containing ~20 adultA. transitella.Hull-split almonds will be placed in a petridish with a light spray of the repellent solution. A video camera placed above the netted cage will be used to record the landings of the moths onto the petridish for 1 hour. Prior to the test, solvent controls will be tested on the same cages first. Using a randomized block design including repellent-treated and control-treated almonds we we will investigate whether significant decrease of landings occur due to any test repellents. Using this assay we expect to test whether these odorants show repellency against agriculturally important moths, and the optimal odor, and concentration for pilot field trials.

Progress 10/01/19 to 09/30/20

Outputs
Target Audience: We have generated data and knowledge for the following target audiences: researchers in agriculture, graduate students, postdoctoral researchers and undergraduate students. Changes/Problems: Nothing Reported What opportunities for training and professional development has the project provided?A graduate student is working on these projects with plants and insects and is learning about developing assays and protocols. How have the results been disseminated to communities of interest? Nothing Reported What do you plan to do during the next reporting period to accomplish the goals?In the next reporting period we will focus on completing Aim 3. The testing of identified repellents on plant phytotoxicity.

Impacts
What was accomplished under these goals? After successfully completing Aim 1, we are well on our way for Aim 2. Aim 2.Behavioral testing of predicted repellentson Drosophila melanogaster for prioritization of spatial repellents. We have tested nearly 50 odorants so far that were identified in Aim 1 for repellency in Drosophila melanogaster. The trap assy being used allows moderate throughput identification of spatial repellents. The Vapor Pressures and mammalian toxicity values for each of the odorants being tested were also computed using a Machine Learning pipeline. Using this approach we have identified 10odorants that show very strong repellency to the flies. We are curently preparing to test the 10- odorants at various doses in alarger spatial arena to identify the best in class ones. Additionally, we are preparing to test these compounds in phytotoxicity assays on plant seedlings.

Publications

  • Type: Book Chapters Status: Published Year Published: 2021 Citation: Stephen T. Ferguson, Anandasankar Ray, Laurence J. Zwiebel. Olfactory genomics of eusociality within the Hymenoptera. Insect Pheromone Biochemistry and Molecular Biology. (Editors Blomquist-Vogt). Pages 507-546
  • Type: Journal Articles Status: Published Year Published: 2020 Citation: Kumar, Arun; Tauxe, Genevieve M; Perry, Sarah; Scott, Christi Ann; Dahanukar, Anupama; Ray, Anandasankar, 2020,Contributions of the Conserved Insect Carbon Dioxide Receptor Subunits to Odor Detection. Cell Reports 31, 107510. https://doi.org/10.1016/j.celrep.2020.03.074


Progress 12/07/18 to 09/30/19

Outputs
Target Audience: We have generated data and knowledge for the following target audiences: researchers in agriculture, graduate students, postdoctoral researchers and undergraduate students. Changes/Problems: Nothing Reported What opportunities for training and professional development has the project provided? Nothing Reported How have the results been disseminated to communities of interest? Nothing Reported What do you plan to do during the next reporting period to accomplish the goals?In the next period we plan to complete the behavioral testing in Drosophila for the volatile predicted repellents and identify the top hits for testing in Agricultural pests.

Impacts
What was accomplished under these goals? Aim 1.Computationally screen 360,000 natural chemicals for predicted repellents and prioritize compounds with high vapor pressure.We were able to utilize the repellency index values for multiple new compounds tested previously, (~100), to improve the selection of a few out of ~5000 DRAGON molecular descriptors, that could best classify activity of the compounds. After identifying the most suitable set of DRAGON descriptors, we trained a Machine Learning algorithm to learn to predict the activity of the training set of repellent compounds. A computational test ws performed to determine the accuracy of the prediction model by a standard approach where one part of the known compounds were set aside as atest set. The model was created using the compounds without the test set, and predictions for repellency were made on the test set chemicals. The true positives and false positives were recorded and used to create an ROC (Receiver Operator Curve). The Area Under the Curve (AUC) was calculated to determine the prediction success rate on the test set. This operation was repeated 100 times with setting aside randomly a different subset of chemicals. When the AUC value for the Machine Learning modelwas >0.85 (or >85% predictive success), it was utilized to screen a >440,000 compound library which contains natural metabolitesfrom plants, microbes and vertebrates. The predicted repellent hits were ranked according to predicted repellency, and predicted Vapor pressure. The highest scoring repellents with the highest vapor presuure values were identified and several were purchased for testing. Behavioral testing for repellency has been initiated in mosquitoes.

Publications

  • Type: Journal Articles Status: Published Year Published: 2019 Citation: Iliano V. Coutinho-Abreu, Kavita Sharma, Liwang Cui, Guiyun Yan and Ray, A. Odorant ligands for the CO2 receptor in two Anopheles vectors of malaria. Scientific Reports. 9: 2549. DOI: 10.1038/s41598-019-39099-0