Recipient Organization
Central State University
1400 Brush Row Rd.
Wilberforce,OH 45384
Performing Department
(N/A)
Non Technical Summary
Honeybees are the most important managed pollinators in U.S. agriculture. To provide the pollination service, honeybee colonies must be transported to orchards and farms by trucks across our country. However, there is a gap in knowledge on the stress of transportation on the behavior and physiology of honeybees. Here, we study how transportation affects honeybee health by investigating the biotic and abiotic stresses in the colonies before, during, and after transportation. We hypothesize that the transportation process and migratory beekeeping can affect bee health at multiple levels, including the immune response, pathogen prevalence, and pesticide resistance. Our project aims to 1) detect the viral prevalence and distribution in transported bees; 2) develop a novel tool to detect honey bee virus using genome editing tools; 3) determine the effects of transportation on changes in Nosema and immune gene expression; and, 4) illustrate the molecular mechanisms of how pesticides affect bees, and 5) use AI and machine learning to develop a new tool to predict the viral stress in workers based on the behavior of workers under viral infections. These studies will discover crucial and novel information on how transportation affects honeybees at individual and colony levels. Our research will help to sustain the survival and health of honeybee populations - ensuring food security and bioeconomy of apiculture. Central State University, Ohio's only public HBCU and 1890 Land-Grant institution in collaborating with Michigan State University to enhance CSU's research capacity and train minority students to join the future workforce for sustainable agriculture.
Animal Health Component
40%
Research Effort Categories
Basic
50%
Applied
40%
Developmental
10%
Goals / Objectives
We hypothesize that the transportation process and migratory beekeeping can affect bee health at multiple levels including the immune response, pathogen prevalence, and pesticide resistance. Our project aims to 1) detect the viral prevalence and distribution in transported bees; 2) develop a novel tool to detect honey bee virus using genome editing tools; 3) determine the effects of transportation on changes in Nosema and immune gene expression; and, 4) illustrate the molecular mechanisms of how pesticides affect bees, and 5) use AI and machine learning to develop a new tool to predict the viral stress in workers based on the behavior of workers under viral infections.
Project Methods
Obj 1. Detecting viral prevalence and distribution in transported bees. The transportation route is from the commercial bee apiary in Boston, GA to almond orchards in Bakersfield, CA (Figure 7) in year 1 and 2. Newly emerged bees from 15 source colonies will be painted marked (each colony a different color), then split and introduced into 30 colonies (N=200 bees per colony, a total of 6,000 bees, with 15 different colors). Fifteen of these colonies will be transported to California for almond pollination, and the other fifteen colonies will stay in GA as the control/stationary colonies. After arriving in CA, all marked bees of 16-day old will be collected in CA to be tested for Obj 1. All marked bees of 8-day old workers will be collected for Obj. 3 and 4B (detoxification enzymes). Control/stationary bees of the same age will be collected in GA. Each colony will be sampled with at least 30 of marked bees (16-day old, about 9-10 days post transportation). Each bee will be extracted for RNA analysis using TRIzol Reagent (Invitrogen, Carlsbad, CA, USA) according to the manufacturer's protocol with the DNase treatment. RNA extractions will be measured by Qubit fluorometer (FisherSci, MA, USA) for quality and quantity and then will be diluted to 200ng/uL.Obj. 2: Direct viral RNA detection using CRISPR-Cas13a A. Design the primers To target DWV, at least two to three different viral genes will be targeted based on the genomic sequencing and characteristics of DWV. RPA amplification primers and LwaCas13a crRNA (28-30 nt) for detecting targeted gene will be designed for high specificity using online tools. Candidate crRNAs should target conserved regions of DWV in various locations. We will minimize the off-targets to otherbee virus genomes. We will also try to improve and optimize reaction conditions. Necessary plasmids will be developed and deposited into the Addgene repository. B. Virus purification and synthesizing We will follow the published protocol (de Miranda et al., 2013) to prepare the viral stock and virus purification. C. Field testing We will use a published protocol to collect hemolymph from honeybee workers' antennae (10 years old) for field testing.Hemolymph samples will be incubated as the nuclease inactivation reaction for 5-25 min at an optimal temperature, then incubated for 5- 10 minutes for viral inactivation, then detection of viral RNA sequence in a reverse transcription reaction (ProtoScript® II Reverse Transcriptase, M0368L, New England BioLabs), recombinase polymerase amplification (RPA) kit. Next the mix will be in another incubation time of several minutes to detect the viral RNA sequence using Cas13. Primers will be ordered from EtonBio.Obj 3. Determine how transportation affects the incidences of Nosema ceranae and gene expression. A. Nosema ceranae incidences in transported and stationary colonies. The marked bees transported from CA to GA (from Obj 1) will be sampled for this objective. The sampling will be done at 8 days after bees arrive at GA to let Nosema ceranae spores to develop (the minimum time is 8 days after spore inoculation). Ten bees will be dissected to check for Nosema ceranae infection status. If infected, the number of spores per bee will be measured by a hemocytometer (Zhu et al. 2014). B. RNA-seq of bees in transported and stationary colonies. For RNA-seq, we will send bee brains (5 brains per trt, 3 reps). Total RNA will be extracted from the tissues using Trizol. RNA quantification and quality assessment will be performed with the Agilent 2100 Bioanalzyer. The RNA will be used for sequencing library preparation with polyA selection and sequenced on DNBSEQ-G400 with paired-end 100bp reads. The samples' raw RNA-seq data will be firstly processed with SOAPnuketo remove low quality reads. Then the filtered data is mapped to the transcriptome sequence of the Amel_HAv3.1 (GCF_003254395.2) reference genome with Bowtie2and calculated the gene expression level by RSEM.Obj 4. Mechanisms of increased sensitivity to pesticides in transported bees. We willuse the mock transportation condition in the lab settings to understand the mechanisms of transportation.A. Is higher pesticide sensitivity due to reduced pollen consumption? Workers from each colony will be divided into two groups, one on a shaker (Eppendorf ThermoMixer R) and one not on a shaker (control) in the same incubator (34.5ºC, R.H. 50%). Both groups will be provided with bee pollen mixed with syrup to have a paste consistency inside 10x75mm glass tubes. Each tube will be weighed before and after consumption (changed every three days). A control tube also weighed to control for moisture loss/gain in the incubator. We have determined the amount of pollen consumed by caged bees infected or not infected with Nosema ceranae in our lab previously. We expect that shaken bees will consume less pollen. Statistics: Pollen consumption between the two treatments will be compared by analysis of variance with days (age of bees) as repeated measures. B. Is the higher pesticide sensitivity due to sleep deprivation? We will use the method used similarly to Klein et al (2010), with study bees tagged with small iron metal discs while control bees with plastic tags and a plexiglass with many small magnets will be used to move across an observation hive at night to disturb the study bees (sleep deprived bees) while leaving the control bees not disturbed. Both groups of bees will then be sampled and tested for pesticide sensitivity. We will use imidacloprid and thiamethoxam, two of the most used neonicotinoid pesticides (Saleem et al 2021) and test bees by feeding them at different pesticide concentrations inside sugar syrup, so the LC50 can bee calculated. Statistics We will test bees from 3 different genetic sources inside the same observation hive so there is replication across workers from different queens.Obj 5: Develop a tool using AI for beekeepers to better monitor bee behavior through course development and teaching AI at CSU. Monitoring intricate social interactions like trophallaxis among honeybees using AI in a controlled laboratory environment involves a multi-step process. Here's a high-level guide on how out team plan to use Python and advanced techniques to develop the tool. 1. Data collection and preprocessing:We will set up an arena at CSU Bee research lab with controlled environment and cameras positioned strategically to capture the 1-day-old adult honeybee worker interactions in a vertically oriented petri dish (100 x 20 mm) with bee wax foundation. Bees will be infected with DWV A honey and pollen dish will be provided for bees as food. We will make sure proper lighting and camera angles for optimal data collection. Each bee in the group will be marked with a different color dot. The camera is Canon PowerShot S5-IS, and the editing tool we will use is MS Movie Make. Libraries like OpenCV will be used to help with background subtraction, object tracking, and feature extraction. Then we will convert video frames into usable data for analysis. 2. Object detection, tracking and behavior recognition: Implement object detection to identify individual honeybees in each frame. Deep learning models like Faster R-CNN can be trained to detect and track honeybees. Then we train a machine learning model to recognize specific social interactions, like trophallaxis, using labeled data. We will also consider using convolutional neural networks (CNNs) for this purpose. Data labeling will involve annotating instances of trophallaxis in the video frames. 3. Data annotation, feature extraction and model training: We will annotate the video frames with the recognized social interactions, marking the frames where trophallaxis or other behaviors occur. This annotated data will be used for model training and evaluation.