Partnership

Backstory

University of Manitoba
Figure 1: Logo of the University of Manitoba

An integral part of a scientific career is collaborative work. Exchanging knowledge, expertise, and parts bring diversity and deeper insights to one’s project. For these reasons, we were interested in a partnership with a team working on a similar topic. In doing so, we were specifically interested in testing the efficacy of other environmentally friendly toxins that target mussels. Hence, we looked for iGEM teams in the Environment track like us. We rapidly got in touch with the University of Manitoba team through Instagram, to see if the interest was reciprocal.

The UManitoba team is located in Winnipeg, Canada. Their project focuses on the spread of Zebra mussels (Dreissena polymorpha) in waterways in the province of Manitoba. Zebra mussels and Quagga mussels share many characteristics. (cf. Project Description).

During our first Zoom meeting, we shared the project rationales. Our projects shared a common topic: the development of toxins to regulate invasive species of mussels’ spread. The UNILausanne’s team worked on the FitD protein, whereas UManitoba worked on two proteins: Act (Aerolysin-related cytotoxic enterotoxin) and a truncated version of FitD.

First zoom of UManitoba and UNILausanne
Figure 2: Our first Zoom with UManitoba team

  1. Act from Aeromonas hydrophila is a cytolytic toxin. It is a secreted protein known to form pores in the membrane of the host, leading to the permeability of the membrane and thus death of the host cell. (“Uniprot”, 2019)

    2. Act 3D structure
    Figure 3: Structure of the proaerolysin monomer as seen in the dimeric soluble form of the toxin (1PRE). Figure taken from Iacovache et al., 2011

  2. The truncated version of FitD was designed with the following rationale: The central part of the Fit toxin (fitD) plays an important role in its toxicity (Maria Péchy-Tarr, 2008). Moreover, from the PSIRED bioinformatics database prediction, most transmembrane domains are located within the core 1300-2200 bps of the gene. Thus, UManitoba decided to truncate most of the N- and C-terminal sequences.

    3. Fit genes
    Figure 4: Organization of the insect toxin locus in P. fluorescens strains CHA0 and Pf-5. The locus was termed fit (for P. fluorescens insecticidal toxin). Figure taken from Péchy-Tarr et al., 2008

Objectives

Our interest was to compare toxins, and test which one would work best to kill mussels. Therefore, as a part of our Quagg’out project, we tested the killing efficacy of our own fitD constructs on mussels and compared it to the killing efficacy of the truncated fitD and act constructs. We performed survival assays on mussels using their constructs, as the iGEM team UManitoba were not allowed to perform experiments on mussels due to a difference in regulation between Canada and Switzerland. In exchange, the UManitoba team quantified the FitD protein expression for us, as they established a protocol to quantify protein expression.

We created a Slack channel following our first meeting and were in constant contact with each other. Updates on the tasks, details or suggestions for our projects were given through general channels weekly such as information on optimized strains for protein expression.

UManitoba's Contribution

We were interested in quantifying the cell's production of the FitD toxin and comparing this to the yields of the truncated fitD and act constructs. Hence, the UManitoba’s team quantified the concentration of FitD produced by our strains using SDS-page.

Material and Methods

  1. 4x Laemmli Sample Buffer
    1. 277.8 mM Tris-Cl, pH 6.8
    2. 44.4 % (v/v) glycerol
    3. 4.4 % SDS
    4. 0.02 % bromophenol blue
  2. 10x SDS-PAGE Running Buffer
    1. 250 mM Tris
    2. 144.4 g/L glycine
    3. 10 g/L SDS
  3. 40 % Acrylamide Mix
  4. 1.5 M Tris pH 8.8
  5. 1 M Tris pH 6.8
  6. 10 % SDS
  7. 10 % APS
  8. TEMED
  9. Fixing solution
    1. 25 % isopropanol
    2. 10 % acetic acid
  10. Staining solution
    1. 2.5 g/L Coomassie Brilliant Blue R-250
    2. 30 % methanol
    3. 10 % acetic acid
  11. Destaining solution
    1. 30 % methanol
    2. 10 % acetic acid

Sample Preparation

  1. Collect 1 mL of OD600 = 1 media (eg. 3 mL of OD600 = 0.33), centrifuge at 6,000 x g for 5 mins
  2. Resuspend in 200 mL of lysis buffer and sonicate at 30% on ice for 6 x ‘5 secs, 5 secs pulse’
  3. Centrifuge at 17,000 x g for 5 mins, the supernatant is the soluble fraction
  4. Resuspend the pellet in 200 mL of lysis buffer + 6 M urea, this is the insoluble fraction
  5. Mix 15 µL of sample with 5 µL 4x Laemmli Sample Buffer

Results

UManitoba's SDS-PAGE
Figure 5: Quantification of FitD production using 1. 5% SDS-page, soluble fraction of cell lysate.

We expect FitD to be present at 300 kDa. We observe on the gel a faint band above 180 kDa. The band appears in P. protegens CHA0 FitD, E. coli pSEVA2313 FitD, and induced E. coli pSEVA234 FitD. It does not appear in the remaining samples. However, it is difficult to assume that this band is FitD as its molecular weight is 300 kDa, and unfortunately the ladder used does not allow for identification of proteins larger than 180 kDa.

Due to time constraints, the iGEM team UManitoba could not further develop a more reliable and robust way to detect FitD. However, we are confident that our cells are successful in producing the FitD toxin, since our strains expressing FitD were successful in killing mussels.(cf. Measurements).

UNILausanne's Contribution

UNILausanne Medhi & Jeremy
Figure 6: Here our team members Jérémy and Mehdi received constructs from UManitoba
UNILausanne Lab Picture
Figure 7: Here are the plasmids we received from UManitoba team

We tested iGEM team UManitoba’s toxins on mussels. The tests were performed on Quagga mussels. After receiving the plasmids, we transformed E. coli BL21 (DE3) strain with them.
We followed the same protocol we applied for Pseudomonas protegens (cf. FitD Experiments). Every condition (treatment or control) used 10 mussels. Consequently, we had a treatment with lysed cells for each strain with one optical density.
We tested:

  • Act toxin induced with IPTG (Act+) (UManitoba)
  • Truncated FitD toxin induced with IPTG (FitD+) (UManitoba)
  • Constitutive fitD (UNILausanne)
  • Constitutive fitG (UNILausanne)

With the controls being:

  • Act toxin non-induced (Act-) (UManitoba)
  • Truncated FitD toxin non-induced (FitD-) (UManitoba)
  • E. coli BL21 (DE3) wild type
  • P. protegens CHA0 wild type
  • Lake water

Act Results

Act toxin results
Figure 8: Lysed E. coli cells engineered to express the Act toxin do not kill quagga mussels. Kaplan-Meier graph showing the probability of mussel survival over time when subjected to 8 * 10^8 cells/mL lysed bacterial sample of different strains. Act - is the non-induced toxin expressed in E. coli, Act + is the induced toxin expressed in E. coli, E. coli wild type is the strain BL21 (DE3), lake water is the control where mussels were subjected to Leman lake water only. Statistically significant differences were determined using a Gehan-Breslow-Wilcoxon test, ** significant for p-value < 0.005; *** < 0.001, ns stands for not significant, n = 10 mussels

For this first experiment, we tested the effect of the Act toxin on quagga mussels survivability. We performed our assay with three controls which consisted in applying to live mussels E. coli BL21 (wt), Lake water and the E. coli cells bearing the Act toxin-circuit but that were non-induced (Act-). The actual treatment assessed here was engineered E. coli cells bearing the Act toxin-circuit that were induced with IPTG (Act +, Figure 8).

The data gathered showed that, while cells bearing the Act circuits exhibited a notable killing activity against mussels compared to the wild-type cells, the non-induced cells (Act-) actually appeared to result in more death than the induced cells (Act+). This surprising observation led us to believe that the death of the mussels were either not attributed to the actual Act toxin but caused by confounding factors unknown to us. Another hypothesis was that the Act circuit was not functioning/regulated as intended. We therefore decided to discard these strains and concluded they were not promising for efficient and reliable killing of invasive mussels.

FitD Results

FitD results of UManitoba
Figure 9: Lysed E. coli cells engineered to express the truncated fitD toxin kills quagga mussels. Kaplan-Meier graph shows probability of mussel survival over time when subjected to different strains, respectively 8 * 10^8 cells/mL lysed bacteria per condition. FitD - is the non-induced toxin expressed in E. coli, FitD + is the induced toxin expressed in E. coli, E. coli wild type is the strain BL21 (DE3), P. protegens CHA0 strain wild type naturally expresses FitD, lake water is the control treated with Leman lake water only. Gehan-Breslow-Wilcoxon test, * significant for p-value < 0,05; **** < 0.0001, ns stands for not significant, n = 10 mussels.

Having set aside the Act strains, we tested in a second experiment the efficacy of killing mussels of the truncated FitD toxin of UManitoba. We performed our assay with four controls which consisted in applying to live mussels E. coli BL21 (wt), Lake water, P. protegens CHA0 (wt) and the E. coli bearing truncated FitD toxin-circuit but that were non-induced (FitD-). The actual treatment assessed here was engineered E. coli cells bearing the truncated FitD toxin-circuit that were induced with IPTG (FitD+, Figure 9).

Our data showed that cells bearing the truncated FitD toxin circuit exhibited a remarkable killing activity against mussels compared to the wild-type cells and the non-induced cells (FitD-, Fig. 9). Indeed, we observed that while lake water, uninduced cells and wild-type E. coli cells did not result in any significant mussel death (i.e. probability of survival > 75% 85 days after treatment), both the P. protegens wild-type cells and the E. coli expressing truncated FitD dramatically killed mussels (i.e. probably of survival reaching 0%). More interestingly, the E. coli cells expressing E. coli expressing truncated FitD killed significantly faster than the wild-type P. protegens cells, indicating that at a concentration of 8 * 10^9 cells, UManitoba’s strain was more efficient in killing mussels than a non-engineered P. protegens strain.

Comparing UManitoba and UNILausanne FitD Results

Results of Unil vs Manitoba OD1
Figure 10: Both lysed E. coli cells engineered to express the truncated fitD toxin and lysed P. protegens overexpressing FitD through the transcriptional activator fitG kill efficiently quagga mussels. Kaplan-Meier graph shows probability of mussel survival over time when subjected to different strains at a concentration of 8 * 10^8 cells/mL. FitD + is the induced toxin expressed in E. coli designed by UManitoba, P .prot FitD (const) is the constitutive FitD toxin expressed in P. protegens designed by UNILausanne, P. prot FitG (const) is the overexpression of fitD through the transcriptional activator fitG expressed in P. protegens designed by UNILausanne, E. coli wild type is the strain BL21 (DE3), P. protegens CHA0 strain wild type naturally expresses FitD, lake water is the control treated with Leman lake water only. Gehan-Breslow-Wilcoxon test, * significant for p-value < 0,05; **** < 0.0001, ns stands for not significant, n = 10 mussels.

For the third experiment, we tested the effect of cells bearing different FitD toxin-circuit on quagga mussels survivability. We were particularly curious to see if our own engineered P. protegens could match UManito’s strain killing activity. We performed our assay with two controls which consisted in applying to live mussels E. coli BL21 (wt) and P. protegens CHA0 wt. The actual treatments assessed here were engineered E. coli cells bearing the truncated FitD toxin-circuit that were induced with IPTG (FitD+, Figure 9), P. protegens cells bearing our FitD toxin-circuit (FitD const, Figure 10) and P. protegens cells bearing the fitG transcriptional activator that overexpresses the FitD toxin-circuit (FitG const, Figure 10).

We were pleased to observe that E. coli cells bearing the truncated FitD toxin-circuits and our P. protegens cells bearing the fitG transcriptional activator that overexpresses the FitD toxin exhibited a remarkable and similar killing activity against mussels compared to their corresponding non-engineered strains. Indeed, our data showed that both our engineered P. protegens strain and UManitoba’s E. coli expressing truncated FitD significantly killed mussels as their probability of survival reached 0% within 35 days post-treatment. Interestingly, our data also suggests that the truncated version of FitD used by UManitoba displays more killing activity than our native FitD in P. protegens, with a significant delay in killing noted for our FitD-expressing strain compared to the former.

Conclusion

In conclusion, our partnership was meaningful on different levels. First, we were able to show that our fitG construct caused toxicity to the mussels, and was faster in disposing of invasive Quagga mussels when compared to P. protegens wild-type cells. When comparing mussel killing activity, we concluded that our FitG construct and UManitoba’s truncated FitD construct have similar efficacy.

Interestingly, collaborating with UManitoba also enabled us to decide on which plasmid to use for further experiments. Moreover, it showed us that E. coli BL21 (DE3) is an organism suitable for FitD expression. In the future, we could consider expressing our construct in E. coli BL21 (DE3) instead of P. protegens and compare its FitD expression to UManitoba’s truncated FitD. Finally, we were able to see that the truncated version of FitD has a similar efficacy on mussels compared to fitG overexpression of FitD. UManitoba also provided us with a very insightful paper from Péchy-Tarr et al., giving us more information about fit genes.These findings will help us and other teams to design experiments to tackle quagga and zebra mussels invasion.

Lastly, while UManitoba’s strain helped us immensely in learning more about our project design and the efficacy of the different versions of FitD to kill quagga mussels, we also provided significant support to UManitoba by testing for them their strain, which they could not do.

We thank UManitoba for this collaboration and all the hard work and efforts! We had an amazing time sharing scientific knowledge with you !

References

  1. UniProt. (2019). Retrieved from Uniprot.org website: https://www.uniprot.org/
  2. Iacovache, I., Degiacomi, M. T., Pernot, L., Ho, S., Schiltz, M., Dal Peraro, M., & van der Goot, F. G. (2011). Dual Chaperone Role of the C-Terminal Propeptide in Folding and Oligomerization of the Pore-Forming Toxin Aerolysin. PLoS Pathogens, 7(7), e1002135. https://doi.org/10.1371/journal.ppat.1002135
  3. Péchy-Tarr, M.; Bruck, D. J.; Maurhofer, M.; Fischer, E.; Vogne, C.; Henkels, M. D.; Donahue, K. M.; Grunder, J.; Loper, J. E.; Keel, C. Molecular Analysis of a Novel Gene Cluster Encoding an Insect Toxin in Plant-Associated Strains ofPseudomonas Fluorescens. Environmental Microbiology 2008, 10 (9), 2368–2386 https://pubmed.ncbi.nlm.nih.gov/18484997/