Results

Introduction

To achieve our objective of engineering a modular microcin expression system to target and kill bacteria plant pathogens, we established the following project pipeline:

As we progressed through this pipeline we obtained significant results that showed the identification of novel microcins, function and effectiveness of our microcin expression system, and pathogenicity of our target pathogens.

Testing Our Modular Microcin Expression System

In order to develop our system, we first needed to take the original expression system from Kim et. al and adapt it to be modular (Figure 1). To confirm this, we cloned Microcin V (MccV), a known and well characterized microcin, into our microcin expressing plasmid and tested for its presence (Figure 2). MccV is natively secreted by E. coli that targets and kills other susceptible E. coli lacking its immunity protein (Parker & Davies, 2022).

Figure 1. The original MccV expression plasmid versus the general assembly schema for our modular microcin expression plasmid. Created with BioRender.com.

As a positive control, we used a strain called E. coli W3110 SK01 that contains microcin plasmid pSKP00 with genes for MccV and a secretion system plasmid pSK01 from Kim et al., 2022. This strain has been shown to successfully secrete MccV against susceptible strains of E. coli. We made a plate with E. coli W3110 as the ‘prey’ lawn and compared our modular microcin expressing assembly against the positive control.

Figure 2: Testing our modular microcin expressing plasmid by making it ssecrete a well-characterized microcin to be compared against a positive control secreting the same microcin. 1) Our modular microcin expressing plasmid coding for MccV and a secretion system plasmid. 2) The positive control strain expressing MccV and secretion system. 3) Our modular microcin expressing plasmid coding for MccV without the secretion system. 4) The positive control with MccV without the secretion system. 5) The empty chassis of our construct. 6) The empty chassis for the positive control

Upon observing a clearing for our microcin secreting construct in comparison to the positive control, we proved that our microcin expressing assembly plasmid is effective at secreting microcins when coupled with the secretion system plasmid.

Characterizing Potential Inducers

Our tests so far have been using constitutive promoters to regulate microcin secretion which might make our experiments miss some results. This is because the chassis has no way to stop production of the toxic protein and thus might be dying before secreting the amount of microcin that can be observable. To expand upon the control of the expression of our microcins, we have opted to develop several regulated plasmids in addition to the constitutive expression plasmids we have been using for testing. We characterized several promoters and regulatory genes sourced from the E. coli 'Marionette" study (Meyer et al., 2019). Firstly, we tested expression of green fluorescent protein (GFP) using these promoters transformed into P. ananatis PNA 97-1R and E. coli DH5α and compared their fluorescence in induced and uninduced cultures. By showing promoter expression in two different chassis, we can show that these promoters are functional across different chassis and thus, can be used as a component in the modular microcin expression plasmid when it is transformed in a biocontrol or E. coli.

Figure 3: Graphs characterizing expression of various promoters based on florescence of GFP for pTet (anhydrotetracycline regulated via TetR), pTac (IPTG regulated via LacI), pVanCC (Vanillic acid regulated via VanR) and pCin (OHC-14 regulated via CinR). Left: Promoters transformed into P. ananatis PNA 97-1R. Right: Promoters transformed into E. coli DH5α.

Based on the clear high expression upon induction of the promoters (Figure 3), we were able to successfully characterize the following promoters based on their GFP fluorescence:

Finding Novel Microcins

To find and characterize novel, putative microcins, we developed a multi-stage pipeline that ranged from computational analysis to laboratory benchwork (Figure 4).

Figure 4: Our pipeline depicting our project to search and test putative, novel microcins from computational output to biological assays.

We began our project with finding putative microcins that could potentially target onion pathogens from the Pantoea genus using a microcin identification pipeline called cinful. After passing Pantoea genomes through cinful, the software provided an output of potential microcin sequences found within those genomes based on a model the software constructs from existing, characterized microcins (Cole et al., 2022). Out of the 1337 Pantoea we screened, we identified 6 unique microcins along with the species they originate from:

cinful also outputs the start and stop position of the microcin sequence in the genome. Using this knowledge, we identified immunity proteins for each of the microcins by looking at open-reading frames using a genome sequence viewer (such as SnapGene) around the microcin based on the following criteria:

For our reference, we constructed a sequence alignment tree (Figure 5) of our chosen microcins to classify the similarities between amino acid sequences. We predicted that the microcin with highest sequence divergence from the rest of the identified microcins would exhibit the most toxicity towards the tested genus. This occurs because native immunity proteins typically only confer resistance to their associated microcin and do not protect against microcins with highly differing sequences.

Figure 5: Nucleotide sequence alignment tree showing degree of similarity between chosen microcins. Each cluster contains the genome that the microcin originates from. This diagram helps determine which microcins chosen from the Pantoea genus are the most divergent to identify the most effective microcin

The sequence alignment chart produced by cinful expresses how similar potential microcin sequences are to each other through distance, with the most similar sequences being closest. The sequence alignment chart for the screened Pantoea species suggested that Mcc02 had the most divergent sequence from the other possible microcins.

To show versatility and modularity of our secretion system, we wanted to test multiple microcins from different origins. Thus, we also screened microcins whose origins are from Erwinia and Xanthomonas genera (Table 1). Due to the lack of time, we could only find immunity proteins for some of these microcins.

Table 1: This table represents amino acid sequences for putative microcins from Erwinia and Xanthomonas genera

We plan to test these putative microcins in the future to further validate our modular microcin expression system and our microcin and immunity protein screening pipeline.

Showing Putative Microcin Effectiveness

To demonstrate if our putative, novel microcins exhibit antibacterial properties against their predicted target, we first conducted a zone of inhibition assay. Since our microcins were selected such that they target bacteria from the Pantoea genus, we conducted zone of inhibition assays with our microcins against pathogenic Pantoea strains. The pathogenic Pantoea tested were:

Since we want our microcins to be able to target and kill pathogenic strains that cause onion center rot, we conducted onion rot assays (check Exhibiting Onion Bacteria Pathogenicity) to prove that the pathogenic strains we use for testing impart the symptoms associated with onion center rot.

Figure 6: Zone of inhibition assay to test putative, novel microcin against their predicted targets, Pantoea. Top: Positive result, ZOI plate with P. ananatis PNA 97-1R as prey and transformed E.coli DH5α as predator. Modular microcin-expressing plasmid Mcc04 with secretion system plasmid (1), modular microcin-expressing plasmid Mcc04 without secretion system plasmid (2) untransformed strain (3). Middle: Two spots showing effect of secretion system on Mcc04 against P. agglomerans PNG 92-11. Bottom: Negative result, Zone of inhibition plate with with P. ananatis PNA 97-1R as prey and transformed E.coli DH5α as predator. Modular microcin-expressing plasmid Mcc06 with secretion system plasmid (1), modular microcin-expressing plasmid Mcc06 without secretion system plasmid (2), untransformed strain (3).

Out of the 6 microcins we selected from the Pantoea genome, we noticed that Mcc04 against Pantoea ananatis PNA 97-1R showed the best measure of clearing out of all the microcin-pathogen combinations during the first round of experimenting. This clearing indicates around the microcin secreting strain that our putative microcin is effective in killing P. ananatis PNA 97-1R. Mcc04 against a lawn of Pantoea agglomerans PNG 92-11 also showed a slight clearing around the microcin secreting strain as well (Figure 6). The clearing around these microcin secreting strains is not as dramatic as the ones shown for MccV. However, it should be restated that the ideal conditions under which these putative microcins are secreted is subject to further testing. We could not replicate the above zones we saw for Mcc04, nor did we have any positive results with other microcins. Thus, we decided to test microcin expression using a more sensitive, quantitative assay: performing growth curves.

We conducted growth curve assays with our modular microcin expression system transformed into pathogenic Pantoea strains such that if we observed an inhibition in growth, we could conclude that the microcin targets that Pantoea strain. To explain how we could obtain a culture of a strain that secretes a protein that can kill itself, we transformed the secretion system plasmid first into the chassis and then transformed the microcin expressing plasmid. To keep the metabolic burden of two plasmids and having a secretion system consistent, our controls contain the secretion system plasmid and the backbone of our modular microcin expressing plasmid. Out of all our putative microcins, we chose Mcc04 and Mcc02 as our candidates for doing a growth curve given Mcc04’s previous success and confidence in Mcc02’s neighboring genomic elements that help determining if the sequence is a microcin or not and its extent of difference in sequence as seen from the sequence alignment tree.

Figure 7: Growth curves for pathogenic Pantoea strains transformed with our modular microcin expression system to observe self-inhibition of growth. The pathogenic Pantoea strains transformed with the modular microcin expressing strain were A) P. agglomerans PNG 92-11 B) P. allii PNA 200-100 C) P. ananatis LMG 2665"

As seen from the graph for P. ananatis LMG 2665, we observed that the growth curve for the microcin expressing strain has markedly downward shift compared to the strain not expressing any microcins (Figure 7). This indicates an inhibition in growth caused by the microcin, providing evidence that our putative, novel microcins exhibit antibacterial properties against their predicted targets while also proving that our microcin screening method is effective.

Showing Putative Microcin and Immunity Protein Effectiveness

Finding the right immunity protein for a microcin is crucial for a chosen chassis to effectively secrete microcins, even if the microcin does not target the chassis since a microcin is a toxic protein regardless (Parker & Davies, 2022). Since Mcc04 has been giving us the best result in targeting and killing pathogenic Pantoea strains, we cloned in the immunity protein we predicted for Mcc04. We conducted a zone of inhibition assay for E.coli containing Mcc04 and its immunity protein against Pantoea ananatis PNA 97-1R, which had previously shown a zone without the immunity protein. We expected the zone formed with the microcin coupled with the immunity protein to be larger and more distinct than the one we found for the strain with only the microcin. This is because the secretion of the microcin should be improved by the presence of an immunity protein.

Figure 8: Zone of Inhibition plate to test our Pantoea microcin and immunity protein assembly against Pantoea ananatis PNA 97-1R. A) E. coli DH5α with the modular Mcc04 and immunity protein expressing plasmid along with the secretion system plasmid B) E. coli DH5α with only microcin backbone and secretion system. C) E. coli DH5α by itself (without microcin, immunity protein or secretion system)

As seen from the plate, there was no distinct zone of inhibition around the microcin and immunity protein expressing strain (Figure 8). This could either be an indication that the immunity protein is not effective or the microcin is not being secreted and diffused properly into the medium. Thus, we decided once again to test the microcin and immunity protein with the growth curve assay as it provides a sensitive, quantitative measurement that would also minimize the problem of the microcin not diffusing properly.

We expected the graph of the strain expressing the immunity protein to show higher growth than the strain expressing only the microcin because the immunity protein confers resistance towards the microcin, making the chassis less susceptible to the microcin’s toxic effects. We transformed P. agglomerans PNG 92-11 with both our modular Mcc04 and immunity protein expressing plasmid and the secretion system plasmid then conducted a 14 hour growth curve along with microcin only expressing plasmid as the positive control (Figure 9).

Figure 9: Two repeats of the growth curve assay done on two different days to show activity of P. agglomerans PNG 92-11 containing the modular microcin and immunity protein expressing assembly. Each faded curve is a replicate of an individual colony of the strain; there are 8 total replicates for each strain. The orange growth curve is the strain containing the microcin and immunity protein assembly and secretion system; blue growth curve is the strain containing only the microcin assembly and secretion system; grey growth curve is the strain containing microcin back bone and secretion system (negative control)

For the first repeat of the growth curve, the microcin only expressing curve shows a downward shift starting from the log phase from the negative control curve proving that the microcin shows effective self-inhibition in the absence of an immunity protein. The microcin and immunity protein expressing curve is shown significantly higher than that for the microcin only expressing curve proving that our immunity protein is effective at imparting resistance against the microcin for the chassis. For the second repeat of the experiment, we got the same results except for the microcin only expressing curve starting its downward shift from the negative control much later. This discrepancy might be explained by the fact that these curves were conducted on two different days with different sets of colonies. With more repeats of this experiment, we can be sure what behavior to expect of Mcc04.

Overall, since the microcin only expressing graph shows a clear downward shift indicating slower growth compared to the control and the strain with the immunity protein, our predicted immunity protein is effective with our putative microcin which also provides proof that our immunity protein screening method is also effective.

All assemblies of our modular microcin expressing system that were made for testing are sequence verified. For details visit the engineering page

Exhibiting Onion Bacteria Pathogenicity

Although the Pantoea strains we used belonged to species typically associated with onion center rot, we still recognized the need to characterize their virulence levels. Upon reviewing the relevant literature, we found that two of our strains were found to have gene clusters associated with the virulence of the rot (Stice et al., 2021). To characterize the pathogenicity of our pathogenic Pantoea strains that we would be testing our microcins with, we conducted onion rot assays. These assays also allowed us to develop a potential in vivo assay to test the effectiveness of our modular microcin expression system against onion pathogens. For the onion rot assay, we injected separate Texas 1015Y Sweet onions with one of our four pathogenic Pantoea strains along with a water control. We incubated all onions at 37°C for 14 days, then observed the rot formed by cutting open each onion at the center. The severity of browning and softening of onion center flesh compared to the water control was observed to evaluate the virulence of each strain.

Figure 10: Photos of the onions cut after 14 days that were inoculated at different optical densities (OD) along with a water control

We observed that P. ananatis PNA 97-1R showed the most virulent pathogenic properties compared to all other strains given that it showed the most browning and softening of flesh after 14 days of inoculation. Based on our microcin testing, one of our putative microcins (Mcc04) is effective at killing the same strain that causes the most pathogenicity out of all the pathogens we are testing. Concluding P. ananatis PNA 97-1R as our most virulent strain, we then planned to characterize the timeline of the rot for this strain as a reference to use when we conduct in-vivo tests of our modular microcin expressing system. We conducted an onion rot progression assay in which multiple onions were inoculated on the same day with P. ananatis PNA 97-1R and were then sequentially cut over 28 days. This allowed us to determine what timeline would be used for our planned in vivo assays.

Figure 11: Photos of different onions inoculated on the same day with P. ananatis PNA 97-1R, E. coli W3110 and water cut on different days.

Using our onion assays, we were able to successfully confirm and characterize the pathogenicity of our Pantoea strains as well as develop a potential framework for testing the effectiveness of our microcin expression system against onion pathogens in vivo.

Conclusion

Based on our goals highlighted earlier, we list the following as what we have achieved for our research project:

Here are our future goals and testing plans for our project:

References

Cole, T. J., Parker, J. K., Feller, A. L., Wilke, C. O., & Davies, B. W. Evidence for Widespread Class II Microcins in Enterobacterales Genomes. Applied and Environmental Microbiology, 2022, 88(23). https://doi.org/10.1128/aem.01486-22

Kim, S-Y., Parker, J. K., Gonzalez-Magaldi, M., Telford, M. S., Leahy, D. J., & Davies, B. W. (2023). Export of Diverse and Bioactive Small Proteins through a Type I Secretion System. Applied and Environmental Microbiology, 89(5), e00335-23. https://doi.org/10.1128/aem.00335-23

Meyer, A. J., Segall-Shapiro, T. H., Glassey, E., Zhang, J., & Voigt, C. A. (2019). Escherichia coli “Marionette” strains with 12 highly optimized small-molecule sensors. Nature Chemical Biology, 15(2), 196-204. https://doi.org/10.1038/s41589-018-0168-3

Parker, J. K., & Davies, B. W. (2022). Microcins reveal natural mechanisms of bacterial manipulation to inform therapeutic development. Microbiology, 168(4), 001175. https://doi.org/10.1099/mic.0.001175

Stice, S. P., Shin, G. Y., de Armas, S., Koirala, S., Galván, G. A., Siri, M. I., Severns, P. M., Coutinho, T., Dutta, B., & Kvitko, B. H. (2021). The Distribution of Onion Virulence Gene Clusters Among Pantoea spp. Frontiers in Plant Science, 12. https://doi.org/10.3389/fpls.2021.643787