Overview of our solution
Sugar beets represent a huge European industry, especially in France (the first European producer) [1]. Our Ionis-Paris team heard about the neonicotinoid’s interdiction in Europe in 2023 and realized that farmers were left without effective solutions for preserving sugar beets. In fact, without neonicotinoids, sugar beet fields become preys to the Beet Yellows Virus (BYV) attacking the beets and consequently reducing the sugar yield [1]. In order to provide help to farmers, we decided to create a biocontrol solution against the BYV, to be sprayed on the sugar beet fields, that would be easy to apply and specific to the virus.
To create the solution, we first needed to establish a way to inhibit the viral infection. That is when we thought of interfering RNA (iRNA) to prevent viral propagation by specifically targeting the virus genome and destroying it [2]. This solution will then not target or kill the aphids which are transmitting the virus, preserving the trophic chains. To avoid developing genetically modified plants, we preferred producing iRNA precursors in bacteria which will then be sprayed on the plants [3, 4]. The plants are then not considered as GMOs[5].
We aim to design and produce long hairpin RNAs (lhRNAs) that will be processed by the cellular machinery of sugar beets into small interfering RNAs (siRNAs) targeting the BYV genome [6]. As RNA is not stable in the environment [7], we will incorporate our interfering RNA precursors into a viral capsid of another plant virus. In this way, they will be protected until their entry into the plant’s cells where DNA release will be facilitated.
Formation of long hairpin RNA
We chose to design long hairpin RNAs [8,9]. These latter are long double stranded RNAs composed of the sense sequence to target, a hairpin loop and the antisense strand of the sequence to target. Compared to the more commonly used short hairpin RNAs (shRNAs), the lhRNAs help to provide a higher stability to our solution. It also increases the chance of being cut by the Dicer complex, responsible of the cutting into siRNAs as it prefers cutting when it recognises loops[10]. Finally, it allows to target a wider part of the virus genome as they will be cut into many different siRNAs, whereas the shRNAs lead to only one, increasing the chance of eliminating the virus [11,12].
To design them, we first thought of the sequences to target in the virus genome to be sure to inhibit the virus reproduction. After some bibliographic research, we decided to focus on the RNA-dependent RNA polymerase (RdRp) and on the sequence coding for the p21 protein. The sequence coding for the RdRp is highly conserved in the virus, meaning that it is not truly mutating with time, and is responsible of synthesising viral progeny in infected cells[13,14]. Inhibiting it will result in a lack of virus replication, and then to its elimination. We kept a sequence of 569 bases (RNA-directed RNA polymerase homolog - sugar beet yellows virus (fragment) [Beet yellows virus] on NCBI, GenBank: X53462.1). The other sequence chosen is the one coding for the p21 protein. This protein acts as a suppressor of double-stranded RNA silencers, and is part of a new family of RNA silencing suppressor present in Closterovirus (the family of the BYV) [15,16]. Eliminating this protein will allow the upholding of our siRNAs, and as well the destruction of a part of the virus genome. We kept a sequence of 408 bases (p21 [Beet yellows virus], on NCBI, GenBank: U71297.1). The next step consisted of creating the loop for the hairpin structure. The loop structure will allow a better detection of the lhRNA by the Dicer complex. It will also confer a better stability to the solution, compared to double-stranded RNA [9]. We needed to incorporate a few bases providing a loop configuration between sense and antisense sequences of the chosen targets. To design the loop, we saw in the literature that for lhRNA designs, we should use loop sequences of around 8 nucleotides for a final size of around 1kb, with a C and a G base at the extremities of the loop for increased stability. It then requires a random mismatch for the loop opening. For construction convenience, we wanted a restriction site in the loop, so that if the sequences were too difficult to synthesize, we could order the sense and the antisense sequences separately and assemble them[10, 17]. We finally chose this one: TCATATG. We tried the in silico-functioning of our construction thanks to the RNAfold WebServer: http://rna.tbi.univie.ac.at//cgi-bin/RNAWebSuite/RNAfold.cgi .
Figure 1: Secondary structure
of the p21 lhRNA
Figure 2: Secondary structure of the p21 lhRNA,
zoom on the loop and the first 21 nucleotides
Finally, we needed to think of a backbone to integrate our sequences, to produce the most RNA as we can from the synthesized sequences. We decided that we would be amplifying our construct into DH5α E. coli bacteria known for plasmid amplification as they have important transformation yields, deficiency in recA1 avoiding clone rearrangements or degradation by recombination and ensuring their stability. These bacteria also have the mutation endA1, leading to deficiency in endonuclease1 which could degrade the plasmidic DNA when extracted [18]. We then chose the pET21a+ plasmid which was already used in one of our school’s laboratories. This plasmid contains an ampicillin resistance cassette, a T7 promoter, and a T7 terminator. The T7 promoter was chosen as it has a high expression capacity and the T7 terminator as it is efficient in stopping transcription [19]. Before trying to build this plasmid in silico, we anticipated the fact that we would protect the long hairpin RNA into a viral capsid [20].
We focused on the capsid of the Tobacco Mosaic Virus (TMV), which is naturally penetrating cells of Amaranthaceae plants which is the family of the sugar beets, so it is not unknown by the plant [21]. The TMV naturally recognizes a sequence called the Origin of Assembly Sequence (OAS), having a particular tertiary conformation [22,23]. Actually, it has loops on specific locations. During encapsidation, the capsid proteins naturally recognize these loops and bind to it. This way they assemble around the length of the RNA sequence [24]. We decided to use this sequence as well and to integrate it into our lhRNAs in order to mimic this encapsidation principle.
This way, we have the backbone to amplify the sequences into the DH5α bacteria, the senses and antisense sequences, the loops, and the OAS.
Figure 3: Secondary structure of the p21 lhRNA with the OAS sequence containing many loops on the topand the loop of the lhRNA at the bottom
We then looked for the restriction enzymes to use to perform the plasmidic assembly of our sequences and the backbone. We decided to use the classical method with restriction enzyme digestions. We finally chose Blp1 and BamH1 and performed the in silico assembly of our sequences for a plasmid with a part of the RdRp sequence and one with the p21 sequence.
We also thought of a control plasmid containing a sequence of a gene coding for the phytoene desaturase. This enzyme is key in the ß-carotene biosynthesis and carotenoids contribute to the stabilization of chlorophyll. Inhibiting it will result in leaf bleaching, or at least to a paler phenotype, 9 days after inoculation[25]. This way, when reaching the inoculation step, we will have a control verifying that our lhRNAs were well incorporated inside the plant and processed into siRNAs. It will allow to verify the feasibility of the whole process, and the use of the interfering RNA technology. We found the sequence of the Phytoene desaturase on NCBI (GenBank: XM_048641752.2)
When the sequences will be assembled and amplified in the DH5α bacteria, we will extract them, do the appropriate controls and transform the solution into E. coli HT115 bacteria, allowing the expression of interfering RNA as they are deficient in RNase-III which is degrading dsRNA[3].
We will then use purification methods to extract them and keep them before inoculation in the plants.
Extraction of coat protein
For the encapsidation, we want to use the coat proteins (CP) from the Tobacco Mosaic Virus (TMV). One way to obtain CP is by using wild type TMV. This technique consists of infecting plants such as Nicotiana benthamiana or Beta vulgaris with the TMV to amplify it in the plants. Then, the goal is to purify the virus and separate the genomic RNA of the virus from their CPs [26]. This would enable us to collect the CPs in larger quantity and to use them to encapsulate our lhRNAs, when the OAS is recognized [20].
Engineering of coat protein
To obtain the CPs, we thought of a second way, by using a plasmid construction. This time, to add a bit of challenge, we decided to perform a Golden Gate assembly. To do so, we looked for the sequence of the TMV linked to the strain referenced NC_001367.1 on NCBI, find the sequence related to the coat protein and ordered it. Then, we chose a T7 promoter, a T7 terminator, a Ribosome Binding Site (RBS), and a backbone from the iGEM database (see on the Parts page). The sequences are send into plasmids and we needed to design primers to extract the fragments of interest. Here are the primers’ sequences that we designed thanks to the sequences found :
Promoter_reverse: 5'- GGT TGG GTC TCG AGT ATT TCT AG -3'
Pr_Promoter_forward: 5'- GCT AAG GTC TCA GGA GTC GTC AC -3'
Pr_Terminator_forward: 5'- ATC CAG GTC TCA GCT TGG TGC TTA -3'
Pr_Terminator_reverse: 5'- TAA TAA ATG GTC TCG AGC GCA A -3'
Pr_RBS forward: 5'- GTT GCC CAA ACT GTA AAG TG -3'
Pr_RBS_reverse: 5'- GTT TAG AGC ATC AAC CAC TGC -3
Pr_CP_reverse: 5'- GGT CTC GAA GCT TAA GTT GCA G -3'
Pr_CP_forward: 5'- GGT CTC AAA TGA TGT CTT ACA GTA TCA C -3'
We designed sequences of primers of 20 to 30 nucleotides having a GC% of 40 to 60% and a difference of less than 5°C in the melting temperatures (Tm) of the forward and reverse sequences.
We designed overhang sequences to add to our sequences and processed to the assembly. The aim is firstly to amplify our sequences by Polymerase Chain Reaction (PCR) before performing a Golden Gate assembly thanks to the enzyme BsaI. We want then to transform the Golden Gate product into E. coli BL21 bacteria which are known for protein expression as they are deficient in proteases Ion and ompT, which allow a better stability of the produced proteins[27]. The next steps consist of purifying and collecting the proteins.
Encapsulation of the long hairpin RNAs
The aim is to assemble the lhRNA inside the coat proteins. To do so, we will need to mix the lhRNAs and the CPs together. Thanks to the OAS sequence mentioned above, the CPs will recognize the assembly sequence and will wrap around the lhRNAs naturally, allowing their encapsulation.
Solution efficiency
In order to verify the efficiency and the efficacy of our solution, we plan to do some reverse-transcriptase quantitative polymerase chain reactions (RT-qPCR) to amplify the Beet Yellows Virus viral RNA and quantify its presence in the sugar beets. The aim is to compare the viral titer in healthy sugar beets, in contaminated sugar beets and in sugar beets treated with our solution. We will conduct the experiment at different time periods pre and post-infection to determine how quickly sugar beets can process the interfering RNAs and assess whether the solution has both curative and preventive effects.
To be able to run the experiment properly, we designed primers for the BYV RdRp (X53462.1) for the Beta vulgaris actin-1, which is supposed to be present in sugar beets. We respected rules of Tm differences, GC% and size. We also ordered primers of housekeeping genes supposed to be always highly expressed in sugar beets and already used primers in the literature for the genes GAPDH and EEF1b2 as controls[28].
BvEF2_E_1 Forward: 5'- AGCTGCGAAAATGGTGAAGT-3'
BvEF2_E_1 Reverse: 5'-AGCGTTGATTTCCCGTGATC-3'
BvGAPDH Forward: 5'-CACCACCGATTACATGACATACA-3'
BvGAPDH Reverse: 5'-GGATCTCCTCTGGGTTCCTG-3'
BYV Reverse: 5'- GGC AGT AAC AAG ATG ATC CG -3'
BYV Forward: 5'- GAC TCC GGG TGA ACT TCT AG -3'
Actin Forward: 5'-CAG GTA TTG TGC TGG ATT CTG G-3'
Actin Reverse: 5'-CAT TAG GTG GTC TGT CAA GTC ACG-3'
To compare the efficiency and efficacy in eliminating the BYV of lhRNAs compared to shRNAs we also designed shRNAs for all of the three sequences targeted (P21, RdRp and PDS). They are containing the OAS as well as sequences of 19 nucleotides taken in the sense sequence of each protein and its antisense sequence (see on the parts page).
Solution stability and degradation
To verify the stability of our solution, we decided that we will run a bunch of tests. For example, by placing our solution in a tube containing soil for several days, we can perform RT-qPCR using primers specific to our lhRNA and compare the results with a control sample containing only the solution. This approach will allow us to determine how quickly our solution is naturally degraded by RNases present in the soil. It will be interesting as well to put our product in different liquid solutions and to see in which one they are best stabilized.
References
[1] Hossain, R., Menzel, W., Lachmann, C. & Varrelmann, M. New insights into virus yellows distribution in Europe and effects of beet yellows virus, beet mild yellowing virus, and beet chlorosis virus on sugar beet yield following field inoculation. Plant Pathol. 70, 584–593 (2021).
[2] Dubrovina, A. S. & Kiselev, K. V. Exogenous RNAs for Gene Regulation and Plant Resistance. Int. J. Mol. Sci. 20, 2282 (2019).
[3] Tenllado, F., Martínez-García, B., Vargas, M. & Díaz-Ruíz, J. R. Crude extracts of bacterially expressed dsRNA can be used to protect plants against virus infections. BMC Biotechnol. 3, (2003).
[4] Gan, D. et al. Bacterially expressed dsRNA protects maize against SCMV infection. Plant Cell Rep. 29, 1261–1268 (2010).
[5] Dalakouras, A. et al. Genetically Modified Organism-Free RNA Interference: Exogenous Application of RNA Molecules in Plants. Plant Physiol. 182, 38–50 (2020).
[6] Vargas, M., Martínez-García, B., Díaz-Ruíz, J. R. & Tenllado, F. Transient expression of homologous hairpin RNA interferes with PVY transmission by aphids. Virol. J. 5, (2008).
[7] Chheda, U. et al. Factors Affecting Stability of RNA – Temperature, Length, Concentration, pH, and Buffering Species. J. Pharm. Sci. 113, 377–385 (2024).
[8] Svoboda, P. & Cara, A. Di. Hairpin RNA: a secondary structure of primary importance. Cell. Mol. Life Sci. 63, 901–908 (2006).
[9] Helliwell, C. A. & Waterhouse, P. M. Constructs and Methods for Hairpin RNA-Mediated Gene Silencing in Plants. in Methods in Enzymology vol. 392 24–35 (Elsevier, 2005).
[10] Bofill-De Ros, X. & Gu, S. Guidelines for the optimal design of miRNA-based shRNAs. Methods 103, 157–166 (2016).
[11] Sano, M., Li, H., Nakanishi, M. & Rossi, J. J. Expression of Long Anti-HIV-1 Hairpin RNAs for the Generation of Multiple siRNAs: Advantages and Limitations. Mol. Ther. 16, 170–177 (2008).
[12] Akbar, S., Wei, Y. & Zhang, M.-Q. RNA Interference: Promising Approach to Combat Plant Viruses. Int. J. Mol. Sci. 23, 5312 (2022).
[13] Pogany, J. & Nagy, P. D. Activation of Tomato Bushy Stunt Virus RNA-Dependent RNA Polymerase by Cellular Heat Shock Protein 70 Is Enhanced by Phospholipids In Vitro. J. Virol. 89, 5714–5723 (2015).
[14] Venkataraman, S., Prasad, B. & Selvarajan, R. RNA Dependent RNA Polymerases: Insights from Structure, Function and Evolution. Viruses 10, 76 (2018).
[15] Ye, K. & Patel, D. J. RNA Silencing Suppressor p21 of Beet Yellows Virus Forms an RNA Binding Octameric Ring Structure. Structure 13, 1375–1384 (2005).
[16] Reed, J. C. et al. Suppressor of RNA silencing encoded by Beet yellows virus. Virology 306, 203–209 (2003).
[17] Herrera-Carrillo, E., Harwig, A. & Berkhout, B. Influence of the loop size and nucleotide composition on AgoshRNA biogenesis and activity. RNA Biol. 14, 1559–1569 (2017).
[18] Anton, B. P. & Raleigh, E. A. Complete Genome Sequence of NEB 5-alpha, a Derivative of Escherichia coli K-12 DH5α. Genome Announc. 4, e01245-16 (2016).
[19] Tabor, S. Expression Using the T7 RNA Polymerase/Promoter System. Curr. Protoc. Mol. Biol. 11, (1990).
[20] Yang, J., Zhang, L., Zhang, C. & Lu, Y. Exploration on the expression and assembly of virus-like particles. Biotechnol. Notes 2, 51–58 (2021).
[21] Jun Yang. Spinacia oleracea proteins with antiviral activity against tobacco mosaic virus. Afr. J. Biotechnol. 11, (2012).
[22] Gaddipati, J. P. & Siegel, A. Study of TMV assembly with heterologous RNA containing the origin-of-assembly sequence. Virology 174, 337–344 (1990).
[23] Saunders, K., Thuenemann, E. C., Peyret, H. & Lomonossoff, G. P. The Tobacco Mosaic Virus Origin of Assembly Sequence is Dispensable for Specific Viral RNA Encapsidation but Necessary for Initiating Assembly at a Single Site. J. Mol. Biol. 434, 167873 (2022).
[24] Smith, M. L. et al. Assembly of trans-encapsidated recombinant viral vectors engineered from Tobacco mosaic virus and Semliki Forest virus and their evaluation as immunogens. Virology 358, 321–333 (2007).
[25] Qin, G. et al. Disruption of phytoene desaturase gene results in albino and dwarf phenotypes in Arabidopsis by impairing chlorophyll, carotenoid, and gibberellin biosynthesis. Cell Res. 17, 471–482 (2007).
[26] Sharma, Y., Srivastava, N. & Bala, K. Neuroprotective ability of TMV coat protein on rat PC-12 cells and it’s in silico study with LRRK2 receptor. Neurol. Res. 40, 1028–1039 (2018).
[27] Ratelade, J. et al. Production of Recombinant Proteins in the lon -Deficient BL21(DE3) Strain of Escherichia coli in the Absence of the DnaK Chaperone. Appl. Environ. Microbiol. 75, 3803–3807 (2009).
[28] Veronika Wetzel et al. The Beta vulgaris-derived resistance gene Rz2 confers broad-spectrum resistance against soilborne sugar beet-infecting viruses from different families by recognizing triple gene block protein 1 (2021). https://bsppjournals.onlinelibrary.wiley.com/doi/10.1111/mpp.13066.
[29] Wetzel, V. et al. The Beta vulgaris ‐derived resistance gene Rz2 confers broad‐spectrum resistance against soilborne sugar beet‐infecting viruses from different families by recognizing triple gene block protein 1. Mol. Plant Pathol. 22, 829–842 (2021).
