Project Description | UNILausanne

The Problem

What are Asian hornets and how did they come to Europe?

The Asian hornet (Vespa velutina), also known as the yellow-legged hornet, is a species native to South-East Asia (Perrard et al. 2014). It was accidentally brought to France in 2004 (Chauzat and Martin 2009), and since then, its population has spread rapidly across the continent, reaching neighbouring countries, including Switzerland, where our university is located.

The Asian hornets' arrival in Switzerland is relatively recent, but it is spreading quickly, raising serious concerns (Figure 1) (Agroscope, n.d.). Currently, there are no reliable methods to stop the growth of V. velutina populations, which is troubling because this invasive species threatens local biodiversity, food production, and the economy.

Figure 1: Observations of V. velutina in Switzerland since 2017: several regions in Switzerland are now invaded by the Asian hornet. Data collected by the Swiss topic centre on fauna (infofauna) (“Info Fauna Carto,” n.d.).

Why exactly are Asian hornets a problem?

The ecosystems and biodiversity are at risk because Asian hornets hunt local insects to feed their larvae. The main prey are bees, including the Western honey bee (Apis mellifera) and wild bees. Other pollinators and insects including flies, wasps, butterflies, moths, spiders, and beetles are also affected (Rome et al. 2021). Bees and other pollinators are crucial for the survival of many wild plant species and crops. In fact, an estimated 75% of crop species that contribute to the global food supply benefit from insect pollination (Klein et al. 2006). Given the reliance of crops and wild plants on pollinators, the Asian hornet’s predation is a threat to food security as well as plant and insect biodiversity.

The problem with V. velutina goes beyond predation. While the larvae feed on protein (i.e. dead insects), adults eat fruits, berries, or other sugar based substances (Nave et al. 2024). This feeding behaviour has been shown to damage fruit crops, particularly grapes. To give an example of how damaging the Asian hornet can be: a survey in Galicia, Spain, found that 83% of farmers experienced some degree of damage to their fruit due to V. velutina (Nave et al. 2024).

When we consider the combined impact of V. velutina on crops—both directly by consuming fruit and indirectly by harming pollinators—along with the financial losses suffered by beekeepers, the estimated annual economic damage caused by this invasive species in France alone reaches 100 million euros (Turchi and Derijard 2018).

How do bees in South-Eastern Asia deal with the Asian hornet?

Answer: Since native South-East Asian bees have lived in proximity with the Asian hornet, they have evolved natural defences against their predator. One of them is "heat-balling" which involves many bees encircling the hornet and furiously flapping their wings, leading to its death because of the heat produced by all the bees’ bodies (Ken et al. 2005). Conversely, since Apis mellifera has been exposed to V. velutina only recently, it did not have time to evolve any defense mechanism.

Why can't we stop the invasion?

The fact that the Asian hornet has been in Europe since 2004, and its spread remains uncontrolled, highlights the inadequacy of current methods to stop it. These methods can be divided into two main categories: broad-spectrum pesticide traps and radio telemetry tracking.

Pesticide traps typically use a sugary substance laced with toxic substances to kill adult Asian hornets. Though they are inexpensive and easy to use for farmers and beekeepers, they cause a significant amount of collateral damage, as other insects also ingest the insecticide. Furthermore, considering that a single nest produces up to 15'000 hornets in a season, killing a few adult hornets does not solve the problem(Turchi and Derijard 2018).

Radiotelemetry works by first spotting an adult Asian hornet, capturing it, and attaching a tracking device to it. Afterward, the hornet is released to return to its nest. Once the approximate location of the nest is revealed, the next step is to pinpoint it by walking around the area with a radio antenna and receiver. After the nest is found, it can be destroyed by using strong pesticides. Since nests are often found high up in trees, professional tree climbers are sometimes needed to assist with the nest's removal (“CABI Tracks down and Destroys Asian Hornet Nest in Switzerland Using Radio Telemetry Technique,” n.d.).

Both methods have clear drawbacks: insecticide traps are harmful to local wildlife and fail to address the root of the problem, as they do not target the nests directly. Radiotelemetry, on the contrary, avoids harming other insects and allows for the destruction of nests. However, it is very costly, time consuming, requiring specialised equipment and trained experts. As a result, deploying radiotelemetry on a large enough scale to effectively slow the spread of the Asian hornet is still unrealistic.

There is a need for a new solution to address this invasive species, and this is where our idea comes in!

Our Idea

A box!

This box is controlled by an image recognition software, which will allow adult hornets to feed on a protein bait. The bait contains bacteria that will specifically target hornets' larvae through RNA interference.

Implementation

We place our box under a beehive
Any insect can fly in, but only Asian hornets are let through, thanks to an image recognition software that only opens the door to the bait if it recognises the Asian hornet.
The Asian hornet accesses a protein bait (i.e. fish) contaminated with engineered Lactococcus lactis, a gut commensal of the insect.
The Asian hornet brings the bait back to the nest and feeds its larvae.
The bacteria is spread around the nest through trophallaxis (the transfer of food or other fluids among members of a community through mouth-to-mouth or anus-to-mouth feeding) (see Federico Cappa Interview)
The bacteria colonise the larvae's gut and release specific shRNA that silences essential larval developmental genes
Many larvae die, significantly harming the colony and preventing its growth. Reducing the quantity of new potential queens would effectively contribute to slowing down the spread of the Asian hornet.

Wouldn't the box attract more Asian hornets to the beehive?

Answer: traps currently in use are often placed near beehives, as Asian hornets are naturally attracted to them due to the bee pheromones. Since hornets are already drawn to the beehive, the box would likely reduce the pressure on the hive by diverting hornets, rather than causing additional harm.

The advantages of our box are :

It is both cost and time effective. Indeed, the box is made with simple, inexpensive materials and the protein bait inside only needs to be changed once a week.
It is easy to use by those directly affected by the Asian hornet (beekeepers, farmers), as it does not require any expertise.
It is specific, leaving other insects unharmed.
It targets larvae rather than adult workers which is potentially a more effective way of preventing the spread of the asian hornet.

Why are we targeting the larvae?

We assume that an approach that focuses on targeting the larvae instead of the adult hornet is optimal primarily because it prevents the next generation of hornets from emerging. This has a more significant long-term impact on reducing the population, compared to pesticide traps, as it directly cuts down the number of future adult hornets that can continue to spread and reproduce (Turchi and Derijard 2018). To test our assumption, we have developed a model which proves that targeting larvae is more efficient than eliminating adults. See our model

How can our system specifically target Asian hornet larvae?

Protein bait

The first layer that ensures that we target the larvae is the use of a protein based vessel for our engineered bacteria. Indeed, only the larvae will ingest the bait, as adults only feed on sugary substances (Nave et al. 2024).

Lactococcus lactis

The second way we ensure specificity to the Asian hornet larvae is through our choice of bacteria. After doing a literature search and a bioinformatic analysis of V. velutina microbiota, we selected L. lactis because it has been found in the larvae’s gut (Cini et al. 2018; Hettiarachchi et al. 2023). After choosing the bacteria, we contacted the research group of Prof. Peter Vandamme from Ghent University working on V. velutina microbiota to ask them for a L. lactis strain isolated directly from a hornet's gut. By using a gut commensal of the Asian hornet, we are hoping to maximise the chances of our engineered bacteria to colonise their gut. Colonisation of the gut would lead to a stable population of bacteria. These bacteria would produce and release larger amounts of shRNA. This would lead to a greater shRNA release and gene silencing, as opposed to a transient bacterial population that would not colonise the gut, and hence only be active for a very limited time.

shRNA sequence

Finally, the third layer of larvae specificity lies in our choice of shRNA sequences produced by our bacteria. We have specifically chosen to silence the chitin synthase gene that is involved in the development of the Asian hornet larvae. This ensures that our shRNAs would affect adult Asian hornets to a lesser extent than they would larvae, if at all.

Design

Bacteria engineering

Why do we want to use shRNA to target the Asian hornet ?

We were looking for a solution that is safe for the environment and other insects, meaning one that specifically targets the Asian hornet. RNA interference (RNAi) through short hairpin RNA (shRNA) delivery could be a great option to achieve this. Simply put, RNAi works by turning off specific genes based on their unique sequence. Since some gene sequences are different in different species, RNAi can be designed to target just the Asian hornet’s larvae by shutting down chitin synthase gene, without harming other insects. RNAi is already being investigated as a way to control pests. For example, to control mosquito population, a mosquito gut commensal was engineered to produce shRNA that will silence larval genes (Ding et al. 2023). We want to apply that same idea to the Asian hornet. To achieve this, we came up with the following plan:

Design

Firstly, we decided to clone our constructs in E. coli first and then transform them into our target bacterium L. lactis, as both can harbour extragenomic DNA in the form of plasmids (Papagianni, Avramidis, and Filioussis 2007).
We needed a plasmid with a specific origin of replication, that could work in both strains. We choose to use pMG36E, which carries a pWV01 origin of replication, that we purchased from Addgene (“Addgene: pMG36E,” n.d.) as a backbone. The plasmid contains an Erythromycin resistance that would allow for easy selection.
We designed a simple construct: promoter-shRNA-terminator. We chose a promoter, PTS-IIC, that was described to work in both L. lactis and E. coli (Ogaugwu et al. 2017). The terminator, rrnb t1-t7te, was taken for the iGEM registry (“Part:BBa K4604043 - Parts.Igem.Org,” n.d.).

Figure 2: Map of the plasmid pMG36E_shRNA_scaffold.

For the shRNA, we designed two inverted repeats corresponding to the target sequence, separated by a small 9nt loop.
As RNA is largely unstable and prone to degradation, we sought to increase its half-life. Within bacterial cells, RNA degradation is carried out mainly by RNAses. Thereby, in parallel to cloning shRNA producing plasmids, we planned to create strains of E. coli and L. lactis lacking the RNAse III, as it could cleave our shRNA (Paddison et al. 2002; Court et al. 2013; Ding et al. 2023) before it would be able to interact with the larval AGO/DICR system.

How did we choose the shRNA sequences and target genes ?

It was crucial to choose the correct genes to target with our shRNAs: we wanted them to be essential enough that silencing them would lead to larvae death, but also to be present in Vespa velutina only, to avoid any off-target effects. We have chosen to target the chitin synthase gene.

Once we had selected the V. velutina genes we wanted to target, we identified regions of 20 to 22 nucleotides within this gene that followed an AA dinucleotide, as it is thought that it maximises RNAi efficiency (“siRNA Design Guidelines | Technical Bulletin #506 | Thermo Fisher Scientific - CH,” n.d.). These regions were then used as inverted repeats, separated by a short flexible loop, to generate our shRNAs.

Finally, the last step was to assess how the shRNA we had designed might affect other species. To do this, we conducted an off-target analysis and found that some of our shRNAs could impact other hornet species, particularly Vespa crabro (the European hornet). While there remained a possibility of affecting the European hornet, we were confident that the use of our image recognition technology would minimise the risk to affect other species.

As we waited for our target genes to be chosen, we began cloning our shRNA expression plasmids using control shRNA sequences; one was given to us by the Schaerli lab (BBa_K5047003), and the other was taken from the 2023 iGEM Estonian team (BBa K4604043) (“Project | Estonia-TUIT - iGEM 2023,” n.d.).

Image Recognition Software

Why do we need our box and the image recognition software?

Ideally, the shRNA would be specific enough to target only V. velutina. However, because the Asian hornet's genome is closely related to that of the European hornet, many developmental genes share similar sequences. This raised the risk of off-target effects, potentially harming the European hornet as well. To avoid causing any harm to the local fauna, we decided to add another layer of protection besides the specific shRNA sequence.

Are European hornets useful?

Answer: European hornets are pollinators and are also important for pest control.

The second layer is an image recognition software. Indeed, as mentioned above, the protein bait containing shRNA - producing bacteria is behind a door that only opens when the image recognition software identifies an Asian hornet. This software was adapted from an already existing one, VespAI (O’Shea-Wheller et al. 2024), that can distinguish Asian and European hornets at a high fidelity rate.

Testing in Yeast

How can we test if our shRNA silences the expression of the targeted gene?

After considering testing in live hornet larvae, we decided against it because we wanted to avoid animal testing, and because they are very difficult to handle in laboratory conditions. (see Karine Monceau Interview). Therefore, we designed a system to test the efficiency of our gene silencing strategy using a yeast system inspired by the 2023 Estonian iGEM team (“Project | Estonia-TUIT - iGEM 2023,” n.d.).

Design

Our testing strategy aimed at creating a yeast strain that expresses the RNA interference machinery, namely the Argonaute protein and the Dicer protein that are part of the RISC complex. These proteins are necessary for RNAi to work and are naturally present in the Asian hornet, but are not present in Saccharomyces cerevisiae. More concretely, we created plasmids containing the different parts necessary for subsequent insertion into the yeast genome (Figure 3). Additionally, to test the efficiency of our shRNA, the yeast would also harbour a plasmid producing our shRNA and another plasmid containing a target of 20 to 22 nucleotides that would be degraded by our shRNA (Figure 4).

Figure 3: Plasmid map of DICR and AGO plasmids separately, and then assembled together into a unique plasmid, for future integration into the yeast genome.

Figure 4: Plasmid maps of the shRNA producing plasmid and the plasmid producing the target sequence followed by a GFP.

In our design of the target plasmid, we made sure to put the target sequence between the coding sequence of the GFP and the transcription terminator. This ensures that an observed decrease in GFP production is purely due to RNA interference and not to transcriptional repression of the reporter. If the target were positioned at the start of the gene, we could not rule out the possibility that the shRNA blocks transcription or translation by binding to the gene or the transcript. However, by positioning the target region at the end of the coding sequence, transcription and translation proceed normally, and the only way the shRNA can function is by triggering the degradation of the transcript.

For the design of the yeast testing system, we chose regulatory regions that work in S. cerevisiae, these included 2 constitutively active promoters: pPGK1 and pTEF1. For the production of the AGO protein we employed pPGK1, a widely used yeast constitutive promoter. For the production of the DICR protein, we sought to use the strongest out of three yeast promoters that were available in the department, to ensure high expression levels of the RISC complex. The parts available to us were: pCYC1 (BBa_K5047043), pADH1 (BBa_K5047042) and pTEF1 (a new basic part characterised in the context of this project, BBa_K5047036).

To compare their relative strengths, we used 3 plasmids (pDA95, pDA120 and pDA122), kindly given to us by the Pelet group, DMF, UNIL. Each plasmid includes one of the constitutive promoters followed by a fluorescent reporter, mCherry.

Figure 5: We measured mCherry fluorescence intensity both at the population level (through a plate reader assay) and at the single cell level (via flow cytometry) in order to choose the strongest promoter for our DICR construct.

Following the assembly of our three constructs (the AGO/DICR plasmid, the shRNA plasmid, and the plasmid containing the GFP+target), we planned to transform them into E. coli before recombining them with the S. cerevisiae genome, in order to get a functional yeast testing system.

triple insertion of plasmids into yeast, with yeast

Figure 6: To test whether our shRNA efficiently silences the target construct, we would need to assess the expression of green fluorescence protein (GFP): if the yeast cells produce GFP, our target gene is not silenced. If the yeast cells show a reduced GFP expression, we can deduce that the mRNA containing our target sequence and the GFP have degraded, which means that our shRNA is effectively silencing its target.

Model

Testing the efficiency of our RNA interference control strategy in silico

During the course of our project, we actively questioned some of the assumptions of our strategy and asked ourselves whether targeting larvae would be more efficient at fighting Asian hornets than targeting adult workers, and whether our RNA interference should be delivered directly in form of a shRNA or through a bacteria producing the shRNA, that could colonise the gut of the Asian hornet and then spread among the colony.

Since answering both questions would require real life experiments that encompass a lot of time and expertise that were not directly available to us, and animal testing, which we wanted to limit, we decided to create a virtual simulation of a bee population under pressure from the Asian hornet. To do that, we built upon an existing mathematical model of bee population (Romero-Leiton et al. 2022), to which we added new terms:

Hornet predation rate on the bee population
Hornet population, modelled after the bee population model, to which we added:

an infection rate by our engineered bacteria,
a cross-infection rate from mature to immature hornets.

Overview of our model and main equations

Equations

We designed a 5-dimensional, continuous model, based on ordinary differential equations. In the model, we simulate 5 sub-populations (eq.1-5):

b, the immature bees (eq. 1),
B, the mature bees (eq. 2),
h, the immature hornets (eq. 3),
H, the mature hornets (eq. 4),
and I, the Asian hornets infected by our engineered bacteria (eq. 5).

We describe the change over time of these sub-populations with parameters such as death rates due to different factors (natural death (μ), stress (σ), predation (k_B), RNAi infection (k, k_h)), maturation rate (ω) and immature production rate (β).

References

Addgene: pMG36E.” n.d. Accessed September 23, 2024. https://www.addgene.org/133853/.
Agroscope. n.d. “Asian Hornet.” Accessed September 12, 2024. https://www.agroscope.admin.ch/agroscope/en/home/themen/nutztiere/bienen/bienenkrankheiten/asiatische-hornisse.html.
“CABI Tracks down and Destroys Asian Hornet Nest in Switzerland Using Radio Telemetry Technique.” n.d. CABI.Org (blog). Accessed September 13, 2024. https://www.cabi.org/news-article/cabi-tracks-down-and-destroys-asian-hornet-nest-in-switzerland-using-radio-telemetry-technique/.
Chauzat, Marie-Pierre, and Stephen Martin. 2009. “A Foreigner in France: The Asian Hornet” 56 (May): 86–91.
Cini, Alessandro, Federico Cappa, Iacopo Petrocelli, Irene Pepiciello, Laura Bortolotti, and Rita Cervo. 2018. “Competition between the Native and the Introduced Hornets Vespa Crabro and Vespa Velutina: A Comparison of Potentially Relevant Life-History Traits.” Ecological Entomology 43 (3): 351–62. https://doi.org/10.1111/een.12507.
Court, Donald L., Jianhua Gan, Yu-He Liang, Gary X. Shaw, Joseph E. Tropea, Nina Costantino, David S. Waugh, and Xinhua Ji. 2013. “RNase III: Genetics and Function; Structure and Mechanism.” Annual Review of Genetics 47: 405–31. https://doi.org/10.1146/annurev-genet-110711-155618.
Ding, Jinjin, Chunlai Cui, Guandong Wang, Ge Wei, Liang Bai, Yifei Li, Peilu Sun, et al. 2023. “Engineered Gut Symbiotic Bacterium-Mediated RNAi for Effective Control of Anopheles Mosquito Larvae.” Microbiology Spectrum 11 (4): e01666-23. https://doi.org/10.1128/spectrum.01666-23.
Hettiarachchi, Amanda, Margo Cnockaert, Marie Joossens, David Laureys, Jessika De Clippeleer, Nicolas J. Vereecken, Denis Michez, Guy Smagghe, Dirk C. de Graaf, and Peter Vandamme. 2023. “Convivina Is a Specialised Core Gut Symbiont of the Invasive Hornet Vespa Velutina.” Insect Molecular Biology 32 (5): 510–27. https://doi.org/10.1111/imb.12847.
“Info Fauna Carto.” n.d. Accessed September 29, 2024. https://lepus.infofauna.ch/carto/58510.
Ken, Tan, H. R. Hepburn, S. E. Radloff, Yu Yusheng, Liu Yiqiu, Zhou Danyin, and P. Neumann. 2005. “Heat-Balling Wasps by Honeybees.” Die Naturwissenschaften 92 (10): 492–95. https://doi.org/10.1007/s00114-005-0026-5.
Klein, Alexandra-Maria, Bernard E Vaissière, James H Cane, Ingolf Steffan-Dewenter, Saul A Cunningham, Claire Kremen, and Teja Tscharntke. 2006. “Importance of Pollinators in Changing Landscapes for World Crops.” Proceedings of the Royal Society B: Biological Sciences 274 (1608): 303–13. https://doi.org/10.1098/rspb.2006.3721.
Nave, Anabela, Joana Godinho, João Fernandes, Ana Isabel Garcia, Maria Amparo Ferreira Golpe, and Manuela Branco. 2024. “Vespa Velutina: A Menace for Western Iberian Fruit Production.” Cogent Food & Agriculture 10 (1): 2313679. https://doi.org/10.1080/23311932.2024.2313679.
Ogaugwu, Christian E., Qiuying Cheng, Annabeth Fieck, Ivy Hurwitz, and Ravi Durvasula. 2017. “Characterization of a Lactococcus Lactis Promoter for Heterologous Protein Production.” Biotechnology Reports 17 (December): 86–92. https://doi.org/10.1016/j.btre.2017.11.010.
O’Shea-Wheller, Thomas A., Andrew Corbett, Juliet L. Osborne, Mario Recker, and Peter J. Kennedy. 2024. “VespAI: A Deep Learning-Based System for the Detection of Invasive Hornets.” Communications Biology 7 (1): 1–11. https://doi.org/10.1038/s42003-024-05979-z.
Paddison, Patrick J., Amy A. Caudy, Emily Bernstein, Gregory J. Hannon, and Douglas S. Conklin. 2002. “Short Hairpin RNAs (shRNAs) Induce Sequence-Specific Silencing in Mammalian Cells.” Genes & Development 16 (8): 948–58. https://doi.org/10.1101/gad.981002.
Papagianni, Maria, Nicholaos Avramidis, and George Filioussis. 2007. “High Efficiency Electrotransformation of Lactococcus Lactis Spp. Lactis Cells Pretreated with Lithium Acetate and Dithiothreitol.” BMC Biotechnology 7 (March): 15. https://doi.org/10.1186/1472-6750-7-15.
“Part:BBa K4604043 - Parts.Igem.Org.” n.d. Accessed September 23, 2024. https://parts.igem.org/Part:BBa_K4604043.
Perrard, Adrien, Mariangela Arca, Quentin Rome, Franck Muller, Jiangli Tan, Sanjaya Bista, Hari Nugroho, et al. 2014. “Geographic Variation of Melanisation Patterns in a Hornet Species: Genetic Differences, Climatic Pressures or Aposematic Constraints?” PLOS ONE 9 (4): e94162. https://doi.org/10.1371/journal.pone.0094162.
“Project | Estonia-TUIT - iGEM 2023.” n.d. Accessed September 24, 2024. https://2023.igem.wiki/estonia-tuit/project.
Rome, Quentin, Adrien Perrard, Franck Muller, Colin Fontaine, Adrien Quilès, Dario Zuccon, and Claire Villemant. 2021. “Not Just Honeybees: Predatory Habits of Vespa Velutina (Hymenoptera: Vespidae) in France.” Annales de La Société Entomologique de France (N.S.) 57 (1): 1–11. https://doi.org/10.1080/00379271.2020.1867005.
Romero-Leiton, Jhoana P., Alejandro Gutierrez, Ivan Felipe Benavides, Oscar E. Molina, and Alejandra Pulgarín. 2022. “An Approach to the Modeling of Honey Bee Colonies.” Web Ecology 22 (1): 7–19. https://doi.org/10.5194/we-22-7-2022.
“siRNA Design Guidelines | Technical Bulletin #506 | Thermo Fisher Scientific - CH.” n.d. Accessed September 23, 2024. https://www.thermofisher.com/ch/en/home/references/ambion-tech-support/rnai-sirna/general-articles/-sirna-design-guidelines.html.
Turchi, L., and B. Derijard. 2018. “Options for the Biological and Physical Control of Vespa Velutina Nigrithorax (Hym.: Vespidae) in Europe: A Review.” Journal of Applied Entomology 142 (6): 553–62. https://doi.org/10.1111/jen.12515.