To produce active nisin, the nisA gene is not the only one that needs to be expressed. NisA has to undergo several post-translational modifications to reach its fully active form. Previous iGEM teams have attempted to produce nisin; however, no team has been able to successfully do so. In 2014, the Groningen team aimed to produce active nisin. In their constructs, they incorporated the nisin immunity genes, nisR and nisK, and two genes required for post-translational modifications of nisin, nisB and nisC. They used fluorescent genes to indicate successful transformation of E. coli. In their results, they were able to transform cells with all gene transformants except for nisB, nisR and nisK. However, the team was unable to successfully express any proteins related to nisin biosynthesis. A few years later, the iGEM 2022 Calgary team attempted to produce another nisin homolog: nisin Q. However, nisQ was the only gene they considered, not accounting for the necessary post-translational modifications required to produce active nisin.
To become fully active nisin, NisA must be post-translationally modified in three ways: dehydration, cyclization, and cleavage. NisB is a dehydratase responsible for the first modification. NisB dehydrates eight serines and threonines to dehydroalanine and dehydrobutyrine, respectively. NisC is a cyclase that is responsible for the second modification. NisC forms five thioether bonds between dehydroamino acids and cysteines, giving nisin its distinct structure. NisP is an extracellular protease responsible for the third modification. NisP cleaves the leader peptide (23 of the 57 amino acids making up NisA), leaving fully modified and active nisin. In the naturally occurring producer, Lactococcus lactis, there are other proteins that play a role in the immunity of the host to nisin and the extracellular transport of NisA, but these proteins do not influence the actual production of nisin. We concluded that NisB and NisC are required to produce nisin. For our project, we wanted to be able to activate fully modified nisin at the site of infection, so we used the leader cleavage site to our advantage. By altering the cleavage site, NisP would not be required. Instead, we utilized the V8 protease, an extracellular protease exclusively produced by Staphylococcus aureus, one of the main pathogens contributing to cellulitis infections. Using the V8 cleavage site, NiSkin will remain in an inactive form until it reaches the site of infection.
To help NiSkin reach the site of infection, we employed the use of the skin-penetrating peptide, TD-1. From literature and consultations with experts, we knew that TD-1 would work best when fused to the protein of interest. However, TD-1 has never been fused to NisA, so the most effective way of linking the two proteins is unknown. Therefore, we designed three experimental fusion proteins. The first construct consists of a truncated leader, with TD-1 and a V8 cleavage site between the leader and the nisin precursor peptide of NisA. The second construct consists of TD-1 fused to the full NisA primary structure with a V8 site between the leader peptide and the nisin precursor. The third construct consists of TD-1 and NisA, with a glycine-serine linker in between the two and a V8 cut site between the leader peptide and the nisin precursor. Along with our fusion protein constructs, we wanted to confirm that we had the correct machinery to produce fully modified nisin, so we made two constructs without TD-1: the first is NisA with a V8 protease site separating the leader peptide and the nisin precursor, and the second is NisA with its original NisP cleavage site. An outline of all of our constructs can be found in Figure 2 on the Experiments Page of our wiki.
We decided to use E. coli BL21 (DE3) as our chassis, with a two-plasmid system. Our plan was to insert nisA and nisB into the open reading frame of the high-copy vector pRSFDuet-1 and insert nisC into the low-copy vector pACYCDuet-1. Once we received the control constructs, we amplified the sequences using PCR. We transformed our plasmids into E. coli DH5α and amplified our plasmids. We confirmed insertion with restriction digests/gel electrophoresis and Sanger sequencing. Then, we transformed our plasmids into E. coli BL21 (DE3) cells to begin protein expression. In total, we created 14 DH5α strains and 17 BL21 (DE3) strains with our genes, which can be seen in Table 1 of the Results page on our wiki.
Once the transformation of pRSFDuet-1-nisA-nisB and pACYCDuet-1-nisC into BL21 (DE3) cells was confirmed, the next step was to express our proteins. Since our vectors use a T7 promoter, we used IPTG induction to over-express our proteins of interest. Our initial protein sample protocol involved incubating a culture with 0.4 mM IPTG for 2 hours, centrifuging 0.5 mL of culture and resuspending in Tris-Cl, adding sample buffer, and boiling the sample for 3 minutes. For our SDS-PAGE gels, we used 13% Tris-glycine gels to test our protein samples. We calculated the predicted molecular weights of our proteins using the amino acid sequence. Our first step was to confirm the expression of NisA, NisB, and NisC on their own. Therefore, using the protocol listed above, we made protein samples for BL21 (DE3) cells containing pRSFDuet-1-nisA, pRSFDuet-1-V8-nisA, pRSFDuet-1-nisB, and pACYCDuet-1-nisC. The first protein that we confirmed expression of was NisC, as we saw a band around the predicted molecular weight of 48 kDa. We were able to confirm expression of V8-NisA, NisA, and NisB, by using an 8-18% Tris-glycine SDS-PAGE, as it allowed us to better visualize bands at the top and bottom of the gels. We saw NisB at its predicted molecular weight of 118 kDa. For V8-NisA and NisA, we did not see bands at the predicted molecular weights of 5.81 kDa, 5.84 kDa, respectively, but we saw clear indication of a protein being induced by IPTG at 11-12 kDa, and we later confirmed through Western Blot. After confirming that the proteins were being expressed on their own, the next step was to express the three other NisA constructs and to co-express all three of our proteins (NisA, NisB, and NisC) in the same chassis. We were able to confirm the expression of our other three constructs: TL-TD1-V8-NisA, TD1-V8-NisA, and TD1-linker-V8-NisA. However, we were having difficulties in co-expressing all three proteins. The results we were getting did not have a clear indication of the presence of bands at the predicted molecular weights, with the induced sample and non-induced examples mostly looking identical.
After remaking protein samples from the whole cell lysate of BL21 (DE3) cells containing the nisA, nisB, and nisC genes multiple times and still getting no results on our gel, we decided to modify our protein sample protocol to get better expression of NisB and NisC. After consulting multiple advisors well-versed in protein expression, we suspected that the short incubation period, the concentration of IPTG being used to induce the samples, the amount of culture we use, and the temperature at which we are inducing the cells could be factors that determine how well NisB and NisC are expressed and how they interact with NisA. The original incubation time we used was 2 hours. We were advised to have a longer incubation time, anywhere from 4 to 6 hours. The concentration we were originally using was 0.4 mM IPTG. Using a lower amount of IPTG could help express NisB and NisC, as it has been shown to improve folding and solubility of proteins. Temperature also has a similar effect; incubating at a lower temperature for a longer duration could improve the folding of NisB and NisC and allow for interactions between the three proteins.
To test which factor(s) affected the expression of our three proteins, we made modifications to our protein sample protocol using V8-NisA expression, changing one variable at a time. First, we tested different incubation times, inducing the cultures with 0.4 mM IPTG and letting samples incubate for 2 hours, 4 hours, 6 hours, and 8 hours before centrifuging. We found that 4 hours was the most effective incubation time. Secondly, we varied the amount of IPTG used to induce, and we found that 0.4 mM had the same effect as 0.1 mM IPTG, so we continued using the same amount as before. Third, we increased the amount of culture we centrifuged from 0.5 mL to 1.5 mL, which gave us larger bands at the predicted molecular weight of our protein. Fourth, we decreased incubation temperature from 37°C to 18°C and extended incubation time to 24 hours. We used BL21 (DE3) cells containing both pRSFDuet-1-V8-nisA-nisB and pACYCDuet-1-nisC. This method was the most effective at co-expressing the three proteins: we saw both distinct bands at the predicted molecular weights of NisB and NisC and an increase in molecular weight of NisA. Such an increase might be due to successful post-translational modifications which would most likely impede movement through the gel.
After designing the initial NisA and TD-1 fusion proteins, we realized that there needed to be some mechanism for isolating the product from the device. To this end, we installed a His6 tag at the product’s N-terminus and conjugated the tag to our fusion protein with a WELQ sequence. This proteolytic cleavage site is cut at the carboxyl side of the glutamine residue. Thus, we planned to purify NisA protein variants via the His6 tag and then remove the amino acids needed for purification, leaving behind pure product.
The pRSFDuet-1 plasmids, which served as vectors for our constructs, contain a translational start codon upstream of our first restriction site. Between the start codon and the restriction site where nisA is inserted is sequence coding for a His6 tag. Thus, we biobricked a sequence coding for the WELQ cut site onto the 5’ end of our nisA constructs so that we are able to cleave the His6 tag after it is no longer needed. Assembly of this His6-WELQ-NisA system required only the two restriction enzymes needed to insert nisA. We verified correct insertion of nisA with restriction site mapping and Sanger sequencing outsourced to Azenta. We verified correct expression of His6-tagged NisA with a Western blot using anti-His6 antibodies.
His6-tagged NisA was purified with nickel-nitrilotriacetic acid beads on a nickel column. Results were verified with an SDS-PAGE gel of the elution fractions. However, the 20 μg/mL yield following purification was too low for downstream applications, considering that yield would further decrease after cleavage of the His6 tag.
Our initial design had successfully purified nisin, but in too small an amount to be useful. Thus, we brainstormed ways to increase yield with our given design. We could increase the volume of cell culture expressing nisin, because more protein would bind the column and more would be eluted; we could increase the bed volume of the column to decrease the loss of nisin in the flow through; we could increase the concentration of imidazole used to elute nisin off the column; finally, we could decrease the elution volume to concentrate the protein that leaves the column.
We repeated purifications of His6-tagged NisA and varied protocol elements. When the cell culture volume was increased 25-fold, we found that no nisin was eluted. The SDS-PAGE gel of the fractionation steps showed that nisin was not present in the high-speed supernatant applied to the column, with the predicted band disappearing after sonication. Fractionation was most likely unsuccessful because the sonicator settings were not altered to account for a denser sample being sonicated. Next, increasing the bed volume during purification actually decreased nisin yield. In the same elution volume as before, less nisin would elute because of a higher binding affinity to the resin. When the imidazole concentration was increased, no change in nisin yield was observed. No nisin had been detected in the final elution with the highest imidazole concentration, so increasing the imidazole concentration of the middle elution at which nisin elutes was unlikely to have an effect. Lastly, we have yet to test whether decreasing the elution volume increases nisin yield.
After optimizing our method to purify His6-tagged nisin, we realized that we would have no way of purifying nisin from the reaction mixture following cleavage of the His6 tag with WELQut protease. Thus, we next plan to order an immobilized version of WELQut protease which can bind the column and cleave nisin on the column such that pure nisin elutes. We could further increase the concentration of our nisin yield with a low molecular weight cut off centrifugal filter that allows solvent to collect at the bottom of the tube while keeping nisin at the top.