Efficient Control of Hazelnut Weevil (Curculio nucum) Population: Introducing CRISPRa Technology

This video provides a brief introduction to the project

Research Background:

Existing significant threat to the hazelnut-growing industry from C. nucum: up to 80 percent of plants can be infested in an orchard [1].

High genetic diversity in Corylus spp. (~500 cultivars) [2,3].

- Long period (~17 years) to release a cultivar with a desired trait using standard breeding methods [3].

- Side effects (beneficial organisms hurt,  human health issues) due to pesticide application [4].

Other factors affecting plant growth and yield (climate change and human activities).

Ultimate Goal:

To find an efficient and eco-friendly solution for controlling hazelnut weevil populations in agricultural systems.

Hypotheses:

1. Existing hazelnut cultivars are vulnerable to the hazelnut weevil threat during the oviposition window.

2. Desired shell hardness achieved by overexpression of cell lignification- related genes with CRISPRa technology significantly improves  plant resistance to the threat.

3. Existing C. avelana cultivars have notably different expression levels of the (target) genes involved in lignin biosynthesis.

4. Modified plants with overexpressed target gene(s) can stay viable despite changes in the specific biosynthesis pathways over the growth period.

Objectives (Hypotheses Testing):

1. To model the natural shell hardening process using the available scientific data for existing Corylus cultivars under the pest threat.

2. To create a mock-up of nut shell formation in plants with overexpressed lignin biosynthesis genes and to compare it to the natural process.

3. To determine expression levels of lignification-related genes in different hazelnut cultivars.

4. To simulate metabolic activity in modified hazelnut plants in silico (within a simulated environment).

Workflow (Steps 1-3):

1-2. (A) Morphological and phenological data collection from the recently published sources [5-8]. (B) Modelling probability outcome (shell hardening) using the logistic regression approach (sigmoid function, in particular) [9,10]. (C) Building  a graph chart to depict the results and compare them between the modelled outputs [11] (Fig. 1).

2-3. Search for the information/data (genes, pathways) related to cell lignification processes in C. avelana.  Accessing publicly available DNA and RNA-seq data [12,13] to perform transcriptomic analysis for different hazelnut cultivars’ samples using bioinformatics: (A) reads mapping (bowtie [14]), (B) Fragments Per Kilobase of transcript per Million mapped reads (FPKM) method (Cufflinks [15]).  Design the efficient guide RNA (gRNA) for vector construction of the target genes [16] (Fig. 2).

4. Running a Reinforcement Learning (RL) agent using the PyTorch library [17] in Colab [18]  to create a simulated environment with the ‘state’ (hazelnut genome) and ‘action’ (modified lignin biosynthesis pathways) definitions using 4CL gene as a template.

The methods and techniques proposed for in-lab work (CRISPR transformation and edited plants evaluation) are also shown in Fig. 1.

Results:

1. The modelling has shown that even the resistant cultivar ‘Segorbe’ (with an increased shell hardening rate) is still at risk of being attacked by the pest for ~ 17d before the shell gets hard enough to prevent it from full penetration.  Such cultivars as ‘Cosford’ stay vulnerable to the pest threat over the whole oviposition period.

2. The plants with overexpressed target (MYB-like) genes achieve desired shell hardness notably (~10 days) earlier than the resistant cultivar (Fig. 3).

3. Based on the FPKM values calculated for the five different hazelnut cultivars with publicly available RNA-seq data, it was observed that (1) both the master switches (MYB46,83) and the activators (MYB58,85) of the genes related to lignin production are present in most of the analyzed samples (Fig. 4).  However, the identified genes were very differently (up to five times) expressed at a time point (Table S1). For the employed sample dataset, the cultivar ‘Tombul’ appears to have increased expression of the lignin biosynthesis genes compared to the others. It’s interesting that a lignin repressor gene (MYB4)  was also notably expressed in this sample at the time point.  Note: The time periods of sampling the plant material for RNA extraction were not provided in the corresponding biosamples records while the number of reads per sample was consistent (~4.5G bases).

For the located MYB-like genes, gRNAs were designed to overexpress them using CRISPRa technology (Fig. 5). The designed gRNAs are found effective for further vector construction (Table S2).

4. The RL-run simulation encountered a 12 day metabolic failure with optimized (1.8) energy input level after five iterations. While desired shell hardness (45.2N) was actually achieved on day 12th, the adenosine triphosphate (ATP) source decreased to (-6.8%) causing the collapse (Table S3).  However, an adjustment (penalized pushing the plant metabolic input beyond the sustainable threshold)  prioritizing shell hardening traits only when cellular resources are high enough to sustain plant growth can be proposed to avoid such a failure. This ensures that a plant remains viable with desired shell hardness and metabolic input balanced. The employed RL approach allowed running a rapid (in-silico) experiment to verify the viability of CRISPRa-modified plants avoiding year-long screening trials.

Conclusions:

The obtained results support the hypotheses stating that both desired shell hardness and formation period can efficiently limit hazelnut weevil population growth and spread in agricultural systems. CRISPRa technology can effectively be used for overexpression of the lignin biosynthesis genes to achieve the desired trait(s) in C. avelana cultivars notably faster than in those obtained by the traditional breeding methods.  Despite the possible metabolic collapse followed by CRISPRa modification, the plants can potentially stay viable during their whole growth period with adjusted cell energy input. The modified plants resistant to the pest will eventually have increased yield and product quality achieved without pesticide use.

The next steps that may be taken are possible as (with enough funding) testing the RL-agent against real conditions, and of course actually putting CRISPRa into action as even as a recent technology it has still been used and field tests may be run after getting all of the paperwork done.

Acknowledgements of Assistance

I would like to thank my family for supporting me throughout this project, my dad for introducing me into the vast field of biology particularly plant biology and plant science. And Mrs. Anderson (my teacher) for suggestions on grammar and other things related to this project.

References:

1.  Shanovich HN, Aukema BH (2022) The biology, ecology, and management of the hazelnut-feeding weevils (Curculio spp.) (Coleoptera: Curculionidae) of the World.  Journal of Integrated Pest Management, 13: 1, 16, https://doi.org/10.1093/jipm/pmac008

2. Salardi-Jost M et al. (2025) Genomic characterization of a global Corylus avellana L. collection: insights into hazelnut ancestry and genetic determinants of production traits. BMC Plant Biol. 25(1): 1738.  https://doi.org/10.1186/s12870-025-07596-2.

3. Mehlenbacher S & Molnar T (2021) Hazelnut Breeding. https://doi.org/10.1002/9781119828235.ch2.

4. Ahmad MF  et al. (2024) Pesticides impact human health and the environment with their mechanisms of action and possible countermeasures. Heliyon. 10(7):e29128.  https://doi.org/10.1016/j.heliyon.2024.e29128.

5. Braga GC et al. (1999) Bio-yield force of hazelnut (Corylus avellana L.) shell. Journal of Food Process Engineering, 22(6): 415–426. https://doi.org/10.1111/j.1745-4530.1999.tb00500.x

6. Guiotto A et al. (2021). Morphological and mechanical characterization of hazelnut shells (Corylus avellana L.). Journal of Food Engineering, 292: 110550. https://doi.org/10.1016/j.jfoodeng.2021.110550

7. Thompson DC et al. (1996) Evaluation of Corylus cornuta for resistance to eastern filbert blight and Morphological Traits. HortScience, 31(4): 673c–673. https://doi.org/10.21273/HORTSCI.31.4.673c

8. van de Kamp T et al. (2014) Three-dimensional reconstructions of weevil snouts reveal complex musculature and mechanical adaptations for drilling. PLoS ONE, 9(1): e86606. https://doi.org/10.1371/journal.pone.0086606

9. Bui LN, Ding Q (2025) Logistic regression modeling: methodological insights and roadmap. Curr Pharm Teach Learn. 17(12):102460. doi: 10.1016/j.cptl.2025.102460.

10. https://www.tinkershop.net/ml/sigmoid_calculator.html

11. Wickham H (2016) ggplot2: Elegant Graphics for Data Analysis. Springer-Verlag New York.

12. GenBank [Internet]. Bethesda (MD): National Library of Medicine (US), National Center for Biotechnology Information; 1982. Available from: https://www.ncbi.nlm.nih.gov/nucleotide/

13. Sequence Read Archive (SRA) [Internet]. Bethesda (MD): National Library of Medicine (US), National Center for Biotechnology Information; 2009. Available from: https://www.ncbi.nlm.nih.gov/sra/

14. Langmead B, Salzberg SL (2012) Fast gapped-read alignment with Bowtie2. Nat Methods. 9(4):357-9. doi: 10.1038/nmeth.1923.

15. Trapnell C, Williams BA, Pertea G et al. (2010) Transcript assembly and quantification by RNA-Seq reveals unannotated transcripts and isoform switching during cell differentiation. Nat Biotechnol. 28(5):511-5. doi: 10.1038/nbt.1621.

16. Concordet JP, Haeussler M (2018) CRISPOR: intuitive guide selection for CRISPR/Cas9 genome editing experiments and screens. Nucleic Acids Res. 46(W1):W242-W245. doi: 10.1093/nar/gky354.

17. https://pytorch.org/

18. https://colab.research.google.com/