Abstract

In this comprehensive project, we engaged in an in-depth exploration of bio-genetically engineered pesticides. Our goal is to create an effective and safe fungal biopesticide. We successfully introduced the insect-specific toxin, LqhIT2, to Metarhizium anisopliae ACCC30104, and it significantly improved the virulence of the fungus towards pests. To find the best suicide switch, we tested different combinations of Pmcl1/Pmcl1(short) and SuperNova/KillerRed. The results showed that Pmcl1(short)-SuperNova was the most effective suicide switch combination.

1. Design

1.1 The fungus: Metarhizium anisopliae

Starting with fungal biopesticide selection, our choice is Metarhizium anisopliae. This fungus is a well-documented pathogen of insects with a broad host range, capable of parasitizing over 200 distinct insect species. Its safety profile was particularly interesting: it poses no harm to humans, other mammals, or the environment at large (Kim et al., 2020). This fungus naturally occurs in soils across the world and has been previously utilized as a biopesticide due to its strong insecticidal properties. Specifically, we planned to focus on M. anisopliae ACCC30104, which exhibited strong pathogenicity.

Figure 1. M. anisopliae ACCC30104 on PDA plate

1.2 Plasmid selection

Resistance of M. anisopliae ACCC30104 to several antifungal agents was tested to determine the selection marker to use.

Problem

According to Inglis et al., 2000, M. anisopliae should be susceptible to glufosinate. So we planned to use glufosinate as its selection marker and chose to use pBARGPE1 as the backbone. However, when we tested M. anisopliae ACCC30104 on PDA plates containing 100 µg/mL and 300 µg/mL glufosinate, it grew well. As a result, we had to test more antifungal agents and change our plasmid choice.

Problem solved

We prepared PDA plates containing 100 µg/mL, and 300 µg/mL G418, 100 µg/mL, and 300 µg/mL glufosinate, 10 µg/mL, 100 µg/mL, and 300 µg/mL benomyl (dissolved in ethanol), 0.1%, 0.2%, and 2% ethanol (control of benomyl group). M. anisopliae spores were transferred to the plates and cultured at 30 ℃. The diameters of colonies were measured on days 0, 3, 5, and 10.

Figure 2. A-F. M. anisopliae ACCC30104 growth on PDA plates on day 5. A. PDA control; B. PDA containing 100 µg/mL G418; C. PDA containing 100 µg/mL glufosinate; D. PDA containing 10 µg/mL benomyl; E. PDA containing 100 µg/mL benomyl; F. PDA containing 300 µg/mL benomyl; G. Growth curve of M. anisopliae ACCC30104 on G418 plates; H. Growth curve of M. anisopliae ACCC30104 on glufosinate plates; I. Growth curve of M. anisopliae ACCC30104 on benomyl plates.

Results showed that our M. anisopliae strain ACCC30104 is only susceptible to benomyl, so we chose to use pBARGPE1-EGFP-BenA as our backbone. pBARGPE1 is a classic filamentous fungus-Escherichia coli shuttle expression vector. Compared to the original pBARGPE1 vector, the pBARGPE1-EGFP-BenA changes the bialaphos resistance gene (BarR) to the benomyl resistance gene (BenA), which is suitable for our strain.

Figure 3. Plasmid map of pBARGPE1-EGFP-BenA.

1.3 The toxin: LqhIT2

To improve the efficiency and effectiveness of M. anisopliae as a pesticide, one strategy is to introduce a safe toxin. We selected LqhIT2, a 61-amino acid-long depressant toxin derived from the scorpion species Leiurus quinquestriatus hebraeus. This toxin displays a high affinity for voltage-gated sodium channels in insects, contributing to its potent insecticidal activity (Karbat et al., 2007). Importantly, LqhIT2 is non-toxic to humans and other mammals, making it a safe choice for our purpose. It is worth noting, however, that the biochemical metabolism of this toxin within crops remains an active area of research, and further exploration will be part of our project scope.

Problem

When submitting the check-in forms, the iGEM Safety and Security Committee raised concerns about LqhIT2. We needed to prove the safety of LqhIT2.

Problem solved

We sought instruction from The Center of Biosafety Research and Strategy at Tianjin University. Prof. Wang from the Center helped to assess the biosafety of LqhIT2. To check the potential of a mutated LqhIT2 becoming toxic to humans, we searched similar sequences of LqhIT2 by running Protein Blast on NCBI. Results showed that genes share a similar sequence with LqhIT2 all encode for insect-specific toxins, which showed the safety of LqhIT2 in our project. See Safety for detailed results. We showed the results to Prof. Wang, and he agreed with us.

To promote the secretion of the toxin, a Metarhizium collagen-like promoter signal sequence (Mcl1ss) should be linked at the 5’ end of the LqhIT2 gene (Peng et al., 2014).

The Pmcl1 is a strictly regulated and hemolymph-inducible promoter from M. anisopliae. Compared to the full-length Pmcl1 (2764bp), a truncated, shorter version of Pmcl1 (1586bp) can lead to a twofold increase in downstream gene expression (Kanjo et al., 2019). This strong hemolymph-inducible promoter was the perfect promoter to be used in our project as it only turns on when hyphae enter the insect body. We planned to put Mcl1ss-LqhIT2 downstream of Pmcl1 (short). This design limited the LqhIT2 strictly inside the insect body and improved the biosafety of our product.

Our next step involved designing the expression vector for LqhIT2. We planned to use EcoRI and NdeI to remove the unwanted EGFP and gpaA promoter of pBARGPE1-EGFP-BenA. Mcl1ss-LqhIT2 would be inserted between the two restriction sites, forming pBARGEP1-Pmcl1(short)-Mcl1ss-LqhIT2-BenA.

Figure 4. A. Map of Mcl1ss-LqhIT2; B. Plasmid map of pBARGEP1-Pmcl1(short)-Mcl1ss-LqhIT2-BenA.

1.4 The suicide switches

Following the preparation of the fungal pathogen and toxin, our attention turned towards the exploration of a suicide switch. This term refers to the introduction of certain genes into target cells, leading to the death of the gene-carrying recipient cells. The primary aim of this strategy was to regulate the reproductive efficiency of specific cells, thereby improving controllability and safety. In particular, we planned to incorporate a light-induced suicide switch into our genetically modified strain of M. anisopliae to prevent its unintended spread into the environment. This switch involves a phototoxic protein, SuperNova, which produces lethal reactive oxygen species (ROS) upon exposure to light. To improve the effectiveness of SuperNova and ROS, a nuclear localization signal (NLS) derived from the SV40 T antigen should be connected to the 3’ end of the SuperNova sequence (Lu et al., 2021). The SV40 NLS should guide the protein to be transported into the cell nucleus, and let ROS attack the most vulnerable genomic DNA (Paardekooper et al., 2019).

The 2016 NYMU_Taipei iGEM Team ligated a KillerRed gene after a full-length Pmcl1 promoter. As described above, Pmcl1 (short) should be three times stronger than the full-length Pmcl1. And Onukwufor et al. have proven that SuperNova can produce three times as much ROS as KillerRed. As a result, theoretically, a Pmcl1 (short) combined with SuperNova should be a stronger suicide switch than the 2016 NYMU_Taipei iGEM Team’s. To verify our hypothesis and to find the strongest suicide switch, we planned to test Pmcl1-SuperNova, Pmcl1(short)-SuperNova, Pmcl1-KillerRed, and Pmcl1(short)-KillerRed.

In terms of designing the expression vector for the suicide switches, we planned to use EcoRI and NdeI to remove the unwanted EGFP and gpaA promoter of pBARGPE1-EGFP-BenA. Suicide switches would be inserted between the two restriction sites, forming pBARGPE1-Pmcl1-SuperNova-SV40-BenA, pBARGPE1-Pmcl1(short)-SuperNova-SV40-BenA, pBARGPE1-Pmcl1-KillerRed-SV40-BenA, pBARGPE1-Pmcl1(short)-KillerRed-SV40-BenA.

Figure 5. A. Map of SuperNova-SV40; B. Map of pBARGPE1-Pmcl1-SuperNova-SV40-BenA; C. Map of pBARGPE1-Pmcl1(short)-SuperNova-SV40-BenA; D. Map of pBARGPE1-Pmcl1-KillerRed-SV40-BenA; E. Map of pBARGPE1-Pmcl1(short)-KillerRed-SV40-BenA.

2. Build

2.1 Vector construction: LqhIT2

Mcl1ss-LqhIT2 (246 bp) was chemically synthesized by GENEWIS (Suzhou, Jiangsu, China). Pmcl1 (short, 1586 bp) was acquired by PCR, with the genomic DNA of M. anisopliae ACCC30104 as the template (Phusion High-Fidelity PCR Master Mix, Thermo Fisher, Waltham, MA, USA). pBARGPE1-EGFP-BenA was cleaved by EcoRI and NdeI (FastDigest, Thermo Fisher, Waltham, MA, USA).

Pmcl1 (short), Mcl1ss-LqhIT2, and the cleaved pBARGPE1-EGFP-BenA vector were connected by Gibson Assembly (ClonExpress Ultra One Step Cloning Kit, Vazyme, Nanjing, Jiangsu, China), forming pBARGEP1-Pmcl1(short)-Mcl1ss-LqhIT2-BenA. The recombinant vector was transformed into M. anisopliae ACCC30104 by protoplast transformation.

Figure 6. Gel electrophoresis of colony PCR products for verification of correct transformation of plasmid pBARGEP1-Pmcl1(short)-Mcl1ss-LqhIT2-BenA into E .coli DH5α.

2.2 Vector construction: the suicide switches

SuperNova-SV40 NLS (747 bp) and KillerRed-SV40 NLS (747 bp) were chemically synthesized by GENEWIS (Suzhou, Jiangsu, China). Full-length Pmcl1 (2764 bp) was acquired by PCR (Phusion High-Fidelity PCR Master Mix, Thermo Fisher, Waltham, MA, USA), with the genomic DNA of M. anisopliae ACCC30104 as the template. pBARGPE1-EGFP-BenA was cleaved by EcoRI and NdeI (FastDigest, Thermo Fisher, Waltham, MA, USA).

To test the effect of Pmcl1, Pmcl1(short), KillerRed, and SuperNova in suicide switch and find the best combination, we planned to construct four recombinant vectors by Gibson Assembly (ClonExpress Ultra One Step Cloning Kit, Vazyme, Nanjing, Jiangsu, China): pBARGPE1-Pmcl1-SuperNova-SV40-BenA, pBARGPE1-Pmcl1(short)-SuperNova-SV40-BenA, pBARGPE1-Pmcl1-KillerRed-SV40-BenA, pBARGPE1-Pmcl1(short)-KillerRed-SV40-BenA.

Problem

Although we succeeded in constructing plasmids containing Pmcl1(short), we could not construct plasmids containing full-length Pmcl1 by Gibson Assembly as expected. The long length of Pmcl1 (2764 bp) might lower the efficiency of assembly. As a result, we had to try another method.

Problem solved

We designed two primers at the 5’ end of the Pmcl1(short, 1586 bp) and linearized the pBARGPE1-Pmcl1(short)-KillerRed-SV40-BenA and pBARGPE1-Pmcl1(short)-SuperNova-SV40-BenA by PCR.

Figure 7. A. Map of linearized pBARGPE1-Pmcl1(short)-SuperNova-SV40-BenA; B. Gel electrophoresis of colony PCR products for verification of correct transformation of plasmid pBARGPE1-Pmcl1-KillerRed-SV40-BenA (1-4) and pBARGPE1-Pmcl1-SuperNova-SV40-BenA (5-8) into E .coli DH5α.

Then, the truncated part of Pmcl1 (1181 bp) was amplified by PCR and connected to the linearized plasmids to restore the full-length Pmcl1. As a result, pBARGPE1-Pmcl1-KillerRed-SV40-BenA and pBARGPE1-Pmcl1-SuperNova-SV40-BenA were constructed based on pBARGPE1-Pmcl1(short)-KillerRed-SV40-BenA and pBARGPE1-Pmcl1(short)-SuperNova-SV40-BenA.

Each recombinant vector was then transformed into M. anisopliae ACCC30104 by protoplast transformation.

3. Test

3.1 Virulence tests

We tested the virulence of the engineered M. anisopliae on the larvae of the greater wax moth (Galleria mellonella). Spore suspensions from the original strain, the strain expressing LqhIT2, and the strain containing Pmcl1(short)-SuperNova were prepared. Different concentrations of spore suspensions (1 × 10⁷, 1 × 10⁶, 1 × 10⁵ spores/mL) were inoculated to different groups. The larvae were divided into ten groups (15 larvae/group) based on different treatments:

a) Group I: uninoculated control larvae;

b) Group II: inoculate larvae with 1 × 10⁷ WT M. anisopliae spores;

c) Group III: inoculate larvae with 1 × 10⁶ WT M. anisopliae spores;

d) Group IV: inoculate larvae with 1 × 10⁵ WT M. anisopliae spores;

e) Group V: inoculate larvae with 1 × 10⁷ LqhIT2 M. anisopliae spores;

f) Group VI: inoculate larvae with 1 × 10⁶ LqhIT2 M. anisopliae spores;

g) Group VII: inoculate larvae with 1 × 10⁵ LqhIT2 M. anisopliae spores;

h) Group VIII: inoculate larvae with 1 × 10⁷ Pmcl1(short)-SuperNova M. anisopliae spores;

i) Group IX: inoculate larvae with 1 × 10⁶ Pmcl1(short)-SuperNova M. anisopliae spores;

j) Group X: inoculate larvae with 1 × 10⁵ Pmcl1(short)-SuperNova M. anisopliae spores;

The larvae and fungi were incubated at 30 °C, and the death numbers of each group were recorded every morning for 15 days.

Figure 8: The larvae of G. mellonella infected by LqhIT2 M. anisopliae. A. Alive on day 1; B. Dead on day 4; C. Hyphae appeared on the surface on day 7; D. Hyphae covered the body on day 10.

Figure 9: A. The survival rate curve of WT groups; B. The survival rate curve of LqhIT2 groups; C. The survival rate curve of Pmcl1(short)-SuperNova groups; D. The survival rate curve of WT, LqhIT2, and Pmcl1(short)-SuperNova groups inoculated with 1 × 10⁷ spores/mL.

The median lethal cell density (LC50) and the median time to death (LT50) were used to evaluate the virulence of each strain. The LqhIT2 strain and the Pmcl1(short)-SuperNova strain had similar LC50 and LT50, indicating that the suicide switch did not influence the virulence of M. anisopliae. On the other hand, the LC50 of the LqhIT2 strain is 25.26-fold lower than that of the WT strain, showing a remarkable increase in virulence. The LT50 of the LqhIT2 strain was also significantly lower than that of the WT strain (p < 0.05). To conclude, the toxin, LqhIT2, significantly increased the virulence of M. anisopliae, and made it more efficient as a fungal biopesticide.

	WT	LqhIT2	Pmcl1(short)-SuperNova
LC50 (day 5)	2.20 × 106	8.71 × 104	2.24 × 106
LT50 (1 × 107)95% CI	4.429 (3.965-4.861)	3.414 (3.024-3.760)	4.4 (3.894-4.864)

Table 1. The median lethal cell density (LC50, spores/mL) on day 5 and the median time to death (LT50, day) under 1 × 10⁷ spores/mL of each strain, calculated using SPSS 26.0.0.2.

3.2 Find the best suicide switch

We tested the suicide switches based on the amounts of spore formation and the spore germination rates. 1 × 10⁷ spores of WT, Pmcl1(short)-LqhIT2, Pmcl1-SuperNova, Pmcl1(short)-SuperNova, Pmcl1-KillerRed, and Pmcl1(short)-KillerRed strains were inoculated to the larvae of the greater wax moth. The Petri dishes containing the pest corpses were placed under sunshine for days until spores formed outside the insect body. Spores were collected by vertexing and the concentrations of spore suspensions were checked under a microscope.

The result showed that the expression of LqhIT2 did not impact spore formation. All the suicide switches worked but with different efficiency. The suicide switch combinations from the strongest to the weakest were Pmcl1(short)-SuperNova, Pmcl1-SuperNova, Pmcl1(short)-KillerRed, and Pmcl1-KillerRed. The Pmcl1(short) generally worked better than Pmcl1, and the SuperNova was stronger than KillerRed. The Pmcl1(short)-SuperNova (p < 0.001), Pmcl1-SuperNova (p < 0.01), and Pmcl1(short)-KillerRed (p < 0.01) strains formed significantly fewer spores than the WT strain. Among all strains, the combination of Pmcl1(short)-SuperNova was the best suicide switch, with a 56.9% decrease in the spore amount. However, this suicide switch is not close to perfect, and it requires further improvement. Melanin synthesized by M. anisopliae absorbs light, and it might help the cells to survive.  

Figure 10. A. The amount of spore formation by each strain on G. mellonella corpses; B. The spore germination rate of each strain in liquid PDB media after 10h. *: p < 0.01; **: p < 0.001.

After being collected, the spores from each strain were resuspended in liquid PDB media at the concentration of 5 × 10⁵ spores/mL. The suspensions were incubated at 30 °C, 180 rpm for 10h to facilitate spore germination.

The spore germination rates of different strains were all high (~92%) and nearly identical. None of the groups showed significant differences (all p > 0.4). The results indicated that our light-induced suicide switches could not influence spore viability after spore formation.

4. Learn

We proved that LqhIT2 could increase the virulence of M. anisopliae and that Pmcl1(short)-SuperNova was an acceptable suicide switch. The results showed that the toxin and the suicide switch worked independently.

Future plans

Our first plan is to incorporate the toxin and the suicide switch into one plasmid, placing them downstream of the same Pmcl1(short) promoter and dividing them by a 2A peptide sequence. Then, the virulence of the fungal biopesticide can be further improved by incorporating other toxins.

In terms of biosafety, Prof. Wang mentioned the concerns about our engineered fungus impacting the environment and gave us suggestions about biosafety tests. In order to stop the uncontrolled spread, improvements should be made to the suicide switch. Prof. Wang suggested that the selection of the control switch should not be overly simplistic, and we should consider multi-level regulatory approaches. First of all, we plan to increase the sensitivity of the light-inducible suicide switch by downregulating the melanin synthases of M. anisopliae. Secondly, we plan to disrupt the RNA-binding proteins that allow it to adapt to the cold. Research has shown that Crp1 and Crp2 protected cells against freezing, and disrupting them would cause the engineered fungi to be killed in winter (Fang & St Leger, 2009).

We plan to test our improved suicide switch systems under various light (high, medium, low), humidity, and temperature conditions. All the tests are planned to be carried out in two stages. Stage one happens in artificial climate chambers. We will introduce soil, Arabidopsis thaliana, and larvae of G. mellonella into the chambers. Stage two happens in experimental fields. We will measure each factor and evaluate the biosafety.

References