We defined the components of microcin expression using the GGA part type categorization described in Leonard et al. 2018 (Figure 3). This allows genetic components of the same part type, such as the microcin and the immunity protein, to be easily interchanged with one another. We made the CvaC15 signal peptide its own part to enable any microcin to be fused to the signal and recognized for secretion. See the Parts Page for more details on how we used these overhangs (Leonard et al. 2018) to define and arrange the different parts in our modular microcin expression system.
Figure 3. The original MccV expression plasmid versus the general assembly schema for our modular microcin expression plasmid. Created with BioRender.com.
We designed primers to amplify the E. coli genes cvaC (MccV) and cvi (immunity protein for MccV) from the original MccV expression plasmid. Because the nucleotide sequence for the CvaC15 signal peptide was too short to design primers for, we designed a pair of complementary oligonucleotides that could be annealed together to form the part. For each part, we ensured that the primers/oligonucleotides added BsmBI and BsaI restriction sites with overhangs that enable the part to fit into our assembly scheme. When cvaC, cvi, and CvaC15 are assembled according to this scheme, they form a full transcriptional unit for expressing MccV.
Using the primers we designed, we performed Polymerase Chain Reactions (PCR) to amplify cvaC and cvi as parts adapted to our GGA standard. To obtain our CvaC15 part, we used a slow-cooling thermocycler protocol to anneal the two complementary single-stranded oligonucleotides we designed together. Once we confirmed that the sizes of each of our parts were correct using gel electrophoresis (Figure 4), we purified them using the Zymo DNA Clean and Concentrate Kit.
We then created two basic parts, one consisting of both cvaC and cvi and the other consisting of CvaC15, by using GGA to clone the parts into a high-copy GFP-dropout entry vector (Leonard et al., 2018) via BsmBI digestion-ligation. Once each basic part was sequence confirmed, we used GGA to assemble these two parts together, along with BTK constitutive promoter (CP25) and terminator (rpoC) parts, into a BTK low-copy backbone via BsaI digestion-ligation (Leonard et al., 2018). The resulting assembly formed a constitutive MccV expression plasmid. Using whole plasmid sequencing, we confirmed that the plasmid we designed in silico had been assembled correctly (Figure 4).
Figure 4. Successful sequencing results for our MccV expression plasmid. The construct that we designed in silico is on top and the sequencing result is on bottom. We used Plasmidsaurus which uses Oxford Nanopore sequencing technology which can have errors in reading stretches of repeated nucleotides. This can result in the appearance of mismatches when aligning the sequences, so we manually examined mismatched regions and confirmed that they were within the backbone and not within the transcriptional unit.
We then electroporated strains of E. coli DH5α and the agricultural biocontrol Pantoea vagans C9-1 with our redesigned MccV expression system along with the secretion system plasmid of the original system (pSK01).
Testing MccV in Our Modular Microcin Expression System
To test if the redesigned MccV microcin expression plasmid functions as intended, we performed Zone of Inhibition (ZOI) assays (Figure 5). As detailed on the Experiments Page , we spotted 10 uL of a concentrated overnight culture of each predator strain onto LB agar plates topped with a layer of soft LB agar containing the prey strain. For MccV, the prey strain was E. coli W3110. The predator strains were either E. coli DH5α or P. vagans. In this experiment, each predator is expressing MccV while the prey is susceptible to this microcin. In Figure 5, the first plate, spot 1 contains our redesigned modular MccV microcin expression plasmid. Spot 2 contains the original MccV expression plasmid from Kim et al.
Figure 5. Result of ZOI assays with E. coli DH5α (left) and P. vagans C9-1 (right) strains. Left: MccV expression plasmid (1), original MccV expression plasmid (2), MccV expression plasmid with pSK01 (3, 4), neither plasmid (5, 6). Right: MccV expression plasmid (1), neither plasmid (2). Created with BioRender.com.
After overnight incubation, we observed that strains containing our redesigned MccV modular microcin expression plasmid displayed visible zones of inhibition.
The production of a ZOI by both strains containing our redesigned MccV modular microcin expression plasmid indicated that MccV, when used in our system, was still expressed, secreted, and effective against its target. This demonstrated that adapting the components of microcin expression to the GGA standard and assembly scheme we created still allows for proper microcin expression and secretion. As such, we learned that our modular microcin expression system is functional.
Additionally, because our system was shown to function in the P. vagans C9-1 strain, we confirmed that our system works outside of E. coli and that P. vagans C9-1 is a suitable chassis for our bioassays. Thus, these results gave us evidence that our system was successfully constructed and functions as intended.
In the hopes of maximizing the likelihood that the Pantoea microcins we selected for further study would be effective, we filtered cinful’s output for microcins that met the following criteria:
The filtered microcins were then aligned using MUSCLE (Madeira et al. 2022) to ensure that the microcins we selected were not duplicates of one another. The alignment indicated that there were six unique microcin sequences, so we chose to work with all six (Figure 6).
Figure 6. MUSCLE alignment of the amino acid sequences of the filtered Pantoea microcins predicted by cinful, with the strain of origin of each microcin indicated on the left. The 90 filtered microcins were clustered into 6 families of unique sequences.
To adapt these microcins to our assembly scheme, we removed the native 15-18 AA-long signal peptide sequence. This signal was found at the beginning of each microcin AA sequence and often ended in a GG, GA, or GS (Parker & Davies 2022). Removing the 15-18 AA-long signal peptide sequence allowed the microcin to fuse with our system’s CvaC15 signal peptide. We then obtained the original nucleotide sequence for each microcin by using information in cinful’s output to locate the microcin sequence in its original genome. We then added BsmBI and BsaI restriction sites that produced the correct overhangs according to our GGA standard (see the Parts Page for more details). We ordered the redesigned microcins as gBlocks from Integrated DNA Technologies (IDT).
After resuspending our Pantoea microcin gBlocks, we created basic parts of each microcin by cloning them into a high-copy GFP-dropout entry vector using GGA with the restriction enzyme BsmBI (Leonard et al., 2018). Once each basic microcin part was sequence confirmed, we used GGA again to assemble each basic microcin part with our CvaC15 part, the CP25 promoter, and the rpoC terminator into a BTK low-copy backbone with the restriction enzyme BsaI (Leonard et al., 2018). The resulting assemblies formed constitutive microcin expression plasmids that followed our assembly scheme. Using whole-plasmid sequencing, we confirmed that the plasmid we designed was assembled correctly (Figure 7).
Figure 7. Successful sequencing results for all 6 Pantoea microcin expression plasmids. The constructs that we designed in-silico is on top and the sequencing result is on the bottom. For reasons described in Figure 4, we manually examined mismatched regions and confirmed that they did not affect the transcriptional unit.
We then electroporated each Pantoea microcin expression plasmid into a strain of E. coli DH5α containing pSK01.
To test whether the Pantoea microcins were effective against any of the 4 pathogenic Pantoea strains we had, we conducted ZOI assays against each pathogen using the E. coli DH5α strains that we transformed in the build stage (Figure 8). For each of the 4 pathogens, we spotted a concentrated overnight culture of each strain (referred to as “predator” strains) onto LB plates topped with a layer of soft agar containing the given pathogen.
Figure 8. Result of ZOI assays with E. coli DH5α (left) and P. vagans C9-1 (right, bottom) strains. Left: Mcc06 (1), microcin expression plasmid with pSK01 (2), neither plasmid (3). Right: Mcc04 (1), microcin expression plasmid with pSK01 (2), neither (3). Bottom:Mcc04 (1), microcin expression plasmid with pSK01 (2), neither (3) backlit for better visualize the presence of the ZOI. Created with BioRender.com.
After overnight incubation, it seemed that a small ZOI was produced by the Mcc04-secreting strain when it was spotted on P. agglomerans PNG 92-11 and P. ananatis PNA 97-1R. However, none of the other microcins produced apparent ZOIs against any of the pathogens.
Since the ZOI generated by Mcc04 against P. agglomerans PNG 92-11 and P. ananatis PNA 97-1R. was very small, it was inconclusive whether the microcin was effective against those two strains. After revisiting microcin testing literature and discussing our results with microcin researcher Jennifer Parker, we learned that observing microcin effectiveness can be influenced by a number of factors (Parker & Davies 2022). Over the course of the rest of the iGEM 2023 season, we went through multiple cycles of modifying our testing procedures to address the following factors:
Microcin Effectiveness Factor
Testing Modifications
Relative concentrations of the microcin secretor and target strains
Tried various combinations of prey and predator concentrations in ZOI
Access to the appropriate outer membrane receptor
Added dipyridyl (DIP), an iron chelator, to our ZOIs to see if increasing availability of iron uptake channels improved ZOI
Inhibition of the microcin secretor strain
Conducted growth curve assays to quantify self-inhibition of growth caused by Mcc04
After first conducting growth curve assays with our E. coli DH5α secretor strains, we chose to also transform our Mcc04 expression plasmid (along with pSK01) into one of our potential target strains: P. agglomerans PNG 92-11. In doing so, we hoped to examine whether Mcc04 caused self-inhibition of growth in the target strain. When we conducted the growth curve with our P. agglomerans PNG 92-11 strain expressing Mcc04, we observed slight growth inhibition (Figure 9).
Figure 9. Growth curve assay for self-inhibition with P. agglomerans PNG 92-11 strains expressing microcin plasmid backbone (yellow), Mcc02 (blue), or Mcc04 (red). All strains also express pSK01. The P. agglomerans PNG 92-11 strain expressing Mcc04 exhibited a visible degree of self-inhibition of growth.
These results led us to believe that Mcc04 may have weak antimicrobial activity against P. agglomerans PNG 92-11. If this is true, self-inhibition of the P. agglomerans strain can be lessened by providing an immunity protein (Parker & Davies 2022). This led us to our most recent cycle which aimed to find and incorporate an immunity protein (IP) into our Mcc04 expression plasmid.
To see if self-inhibition of the target strain caused by Mcc04 could be reduced by the presence of an immunity protein (IP), we decided to incorporate an IP into our Mcc04 expression plasmid. To find a likely IP for Mcc04, we manually searched the genome sequence of the strain of origin for open reading frames near Mcc04 and its native secretion system components (Parker & Davies 2022) (Figure 10). See the Results Page for more on how we found an IP candidate.
Figure 10. Location of the immunity protein candidate for Mcc04 in the genome of a strain of origin. Created with BioRender.com.
To adapt the candidate to our assembly scheme, we designed a DNA sequence that consisted of Mcc04, its IP, and a ribosome binding site (RBS) (Figure 11) with BsmBI and BsaI restriction sites that produced the correct overhangs according to our GGA standard (Figure 3) (see the Parts Page for more details). We then had the final sequence synthesized commercially as a Gene Fragment by Twist Biosciences.
Figure 11. Mcc04 expression plasmid versus Mcc04 + IP expression part. Created with BioRender.com.
After resuspending our Mcc04+IP Gene Fragment, we used GGA with the restriction enzyme BsmBI to clone it into a high-copy GFP-dropout entry vector to create a basic part (Leonard et al., 2018). Once the basic part was sequence confirmed, we used GGA with the restriction enzyme BsaI to assemble the Mcc04+IP basic part with our CvaC15 part, the CP25 promoter, and the rpoC terminator, into a BTK low-copy backbone (Leonard et al., 2018). This assembly resulted in a microcin expression plasmid that expressed both Mcc04 and its IP. Using whole plasmid sequencing, we confirmed that the plasmid we designed in silico had been assembled correctly (Figure 12).
Figure 12. Successful sequencing results for Mcc04 with its immunity expression plasmid. The construct that we designed in silico is on top and the sequencing result is on the bottom.
We then electroporated the new plasmid into strains of E. coli DH5α, E. coli W3110, and P. agglomerans PNG 92-11 containing pSK01.
We conducted 14-hour growth curve assays to compare the growth of the P. agglomerans PNG 92-11 Mcc04-secretor strains both with and without the IP (Figure 13). This allowed us to observe whether the growth inhibition of P. agglomerans PNG 92-11 caused by Mcc04 on its own was remedied by providing immunity. This revealed whether the inhibition caused by Mcc04 (on its own) was due to antimicrobial activity of the microcin.
Figure 13. Two replicate growth curves with P. agglomerans PNG 92-11 strains containing either our Mcc04 + IP expression plasmid (Gray), the Mcc04 expression plasmid without IP (Orange), or just the backbone of our modular microcin expression system (Blue). Each strain also contained pSK01. Averages of each time point are shown in bold while individual time points are grayed out. Time points were taken every half hour.
The results of repeated growth curve assays indicated that the growth of P. agglomerans PNG 92-11 secreting Mcc04 was less inhibited when the IP was expressed.
The growth curve results suggested that expression of the IP lessened inhibition of the Mcc04 secretor strain (Figure 13). This suggested that Mcc04 has promising antimicrobial activity against P. agglomerans PNG 92-11. As such, we were able to learn that cinful can be used to predict microcins that have antimicrobial activity against a pathogen of interest. However, further testing and characterization are needed to fully confirm this.
Additionally, the testing and assembly design changes we made allowed us to devise an improved process for finding and testing microcins. We applied this improved process to find microcins against pathogenic strains of Erwinia amylovora and Xanthomonas. We cloned and assembled these microcins into microcin expression plasmids. However, we have yet to fully test them. See the Results Page for more information on these microcins.
Because the original plasmids from Meyer et al. 2019 only differed in the promoter and regulatory gene being used, we designed two pairs of universal primers to amplify the promoter and regulatory gene of each system. The universal primers captured an RBS and a hammerhead ribozyme for each promoter and a terminator sequence upstream of each regulator. We ensured that the primers added BsmBI and BsaI restriction sites that produce overhangs in accordance with our GGA standard (Figure 14) (see the Parts Page for more details). This way, the promoters would be compatible with our assembly scheme and therefore form versions of our microcin expression system under inducible control.
Figure 14. Inducible promoters and their regulatory gene parts defined according to our assembly scheme.
Using the primers we designed, we performed PCR to amplify Ptet, Ptac, PluxB, PcymRC, pBAD, PvanCC, Pcin, and their corresponding regulatory genes: tetR, lacI, luxR, cymR, araC and araE, vanR, and cinR, respectively. Once we confirmed that the sizes of each of our parts were correct using gel electrophoresis, we purified them using the Zymo DNA Clean and Concentrate Kit.
We then created basic parts of each promoter and regulator by cloning each of them into a high-copy GFP-dropout entry vector (Leonard et al., 2018) using GGA with the restriction enzyme BsmBI. Once each basic part was sequence confirmed, we assembled Ptet, Ptac, PvanCC, and Pcin each with a BTK GFP coding sequence part and their corresponding regulator using GGA. We chose not to create assemblies with the other inducible promoters because we were unable to obtain the inducer molecules for them.
These assemblies resulted in GFP-expression plasmids that enabled us to examine the performance of the inducible promoters. Using whole plasmid sequencing, we confirmed that the plasmids we designed in silico had all been assembled correctly (Figure 15).
Figure 15. Successful sequencing results for GFP assemblies with Ptac (A), PvanCC (B), Pcin (C), and Ptet (D) and their respective regulatory proteins. The construct that we designed in silico is on top and the sequencing result is on the bottom. Created with BioRender.com.
We then electroporated each plasmid into the strains of P. agglomerans PNG 92-11 and E. coli DH5α containing pSK01 we created while pursuing Objective 2. This resulted in a strain that utilized the new inducible promoters while maintaining similar expression to the full two-plasmid system we are using.
We performed a fluorescence assay on our strains containing the Ptet, PvanCC, and Pcin-controlled GFP-expression plasmids and pSK01 (Figure 16). We grew induced and uninduced cultures of the strains overnight and took a single endpoint fluorescence reading for each culture. When taking our readings, we set excitation and emission to 485 nm and 535 nm, respectively.
Figure 16. GFP fluorescence in E. coli DH5α and P. agglomerans PNG 92-11 under the inducible promoters Ptet, Ptac, PvanCC, and Pcin. Ptac was not tested in P. agglomerans PNG 92-11 due to time. Induced (+) cultures exhibited much greater fluorescence than uninduced ones, indicating that the promoters have a clear “on/off” state. Created with BioRender.com.
Our endpoint readings indicated that the induced and uninduced states of each promoter caused significantly different levels of GFP expression.
Through our assay results, we demonstrated that the inducible systems with Ptet, Ptac, PvanCC, Pcin, and their corresponding regulators were still functional when adapted to our assembly scheme and system. We found that the inducible systems were also functional in P. agglomerans PNG 92-11. This indicated that the system is likely also functional in Pantoea strains. From this, we learned that these inducible promoters should enable us to control microcin expression in our system. As such, we chose to make new assemblies with MccV and Mcc04 under inducible expression using these promoters. While we successfully assembled these inducible microcin constructs, we were unable to test them in ZOI or growth curve assays due to lack of time.