AmilCP results

Subproject Achievements

---

Successes

• Found mutations that give amilCP new colors
• Developed a quantitative way of assessing maturation times of chromoproteins
• Implemented and tested Superfolder mutations’ maturation times

Failures

• Were unable to get the random mutagenesis to work
• The Turbofolder mutations were not successfully implemented

First maturation assay

---

The purpose of the first round of maturation measurements was to test out the maturation assay and to assess the maturation time of the green amilCP before introducing the mutations. We made restreaks of the green amilCP (1 and 2), amilCP, and mRFP1 on three agar plates (Figure 1), and let them grow at 37°C for 16 hours. They grew under anaerobic conditions by sealing in a bag purged with nitrogen gas (Bao et al. 2020). During this time, the chromoproteins will be expressed but since oxidation of the chromophore is required for the maturation, no color will develop at this stage. When finally exposing the plates to oxygen, the chromoproteins will start maturing simultaneously. Pictures were taken of the plates regularly over the course of 30 hours at room temperature to monitor the color development (Figure 2; complete compilation of pictures in Appendix A1.1).

From these pictures, the maturation half-times of the chromoproteins can be roughly estimated by eye (Table 1). The maturation half-time is defined as the time it takes for the color intensity of the proteins to reach half of its maximum value.

mRFP1 seems to mature the fastest with a maturation half-time of about 2.5 hours, whereas the green amilCP matures the slowest with maturation half-time of about 7 hours. However, the qualitative way of estimating the maturation time by eye is problematic and is heavily dependent on the bias of the person analyzing the plates. The quality of the pictures is also crucial for a fair assessment of the maturation times. Even though all pictures were taken in the same room under the same conditions (light table on and ceiling lamps off), the brightness of the pictures varies a lot. It turned out that the light table was flickering, making some pictures brighter and some darker than the others. The varying brightness of the pictures naturally makes it more difficult to estimate the maturation rates. Furthermore, it appears that a lot of color development occurred during night time when no one was in the lab to take pictures of the plates (between hour 7 and 23.5). It is thus hard to say exactly when the proteins were fully matured.

Although this initial testing of the maturation assay may not have been so fruitful in terms of good and reliable data, it did highlight the difficulties of the assay and made us realize that we needed make a couple of adjustments before the next round of maturation measurements:

• Use a software program, OpenCv, to quantitatively determine the maturation time of the samples in order to eliminate the bias when analyzing the pictures
• Program the camera to take pictures automatically. In that way, we could take pictures more frequently and also during night time when no one is in the lab
• Replace the light table for another light source that does not flicker
• Change the settings so the camera doesn't auto adjust

Why do amilCP green 1 and 2 look so different?

The color intensity of amilCP green 1 and 2 is strikingly different despite being thought to be expressed from identical plasmids in the same E.coli strain. The fact that the green mutants were made in the same way (N175I, not GFP-numbering; confirmed by sequencing of the amilCP gene), raised suspicions that they might be expressed in different strains. We performed the transformation of 1 into DH5-⍺ ourselves since it was gifted to us as plasmid DNA, whereas 2 was transformed by previous Uppsala University students several years ago. To address our suspicions, the plasmids of amilCP green 2 were extracted and transformed into DH5-⍺. After restreaking both mutants on the same agar plate, they appeared indistinguishable (Figure 3). From here on out, we treated the two mutants as equals.

Color change mutations

---

The results of the sequencing confirmed successful and correct substitutions at position 157 for all the specifically designed primers (Figure 13). The glutamine mutation (Gln157) however also resulted in a deletion of the asparagine at position 156.

Sequencing of the NNN157 mutants showed that out of the 6 sequenced colonies, A and B were Asp mutations, giving a pink color that already was discovered (Figure 13). C proved to be Asn, the original blue amilCP whilst the purple looking mutant D was Ala. The other purple mutant, E, was Gly. The final mutant, F, was Glu which gave a yellow coloring.

In total we got 8 different amino acids at position 157 to give off a color. We had the already known original amilCP (Asn), as well as the green version made with Ile and the pink version made with Asp. Aside from these, we found purple (Ala), light purple (Gly) and yellow (Glu) colors (Figures 8 and 9) which was quantitatively assessed through absorbance measurements (Figures 10, 11 and 12). We also saw indications of a gray-looking mutant (His) and a yellowish orange mutant (Gln), but we didn’t have enough time to characterize these, and thus don’t have any good pictures of them. This result reaffirms the significance of position 157 for the absorption of the chromophore.

An interesting observation made from the Gln mutation is that a deletion right next to this important residue position doesn’t seem to disrupt the overall structure and function of the protein, as seen by the bacteria still giving off color. Running a MSA showed that the region isn’t conserved over the GFP-family. We did not have time to try introducing the same substitution again but without the deletion, but it would be interesting to see if the color differs depending on if N156 is deleted.

Another interesting observation is that the asparagine we got through random primers, where the codon was GAT, seems to differ in its coloring compared to the asparagine we got through specific primers, which was GAC (the same as Team Sydney’s part BBa_K1996003). These absorbance spectra can be found in the Appendix Figure 2.1 and 2.2.

Based on the boxplot of the maturation halftimes (Figure 14), the green amilCP is the fastest maturer among the amilCP color mutants, with a maturation halftime near that of mRFP1. This is not in accordance with our other observations by eye the first maturation assay or in the superfold assay. Moreover, comparing our results from this run with the data reported by Liljeruhm et al. (2018) further challenges the accuracy of our results. Liljeruhm et al. reports a maturation time of 22 minutes for mRPF1 and 54 minutes for amilCP (liquid cultures of E.coli, incubating at 37°C). They also claim that the color development of amilCP green is slow (although provided no half time estimates). Since external factors such as temperature and the oxygen supply affects maturation, it is not fair to compare our half times values to each other (room temperature vs. 37°C; liquid culture vs. colonies on plate). However, the fact that our results indicate that the green mutant matures significantly faster than the blue amilCP (NNN 157 C (c) in Figure 14) and that its halftime is comparable to that of mRFP1 is somewhat puzzling. Looking at the scatter plots, the green amilCP ones are a bit more irregular in their curves, however they still follow the same overall pattern as the other measured samples (see Appendix A4.5 and A4.6). There are a few possibilities for why we got this result:

• The result is accurate, and our estimation by eye is impacted by a bias towards stronger colors when it comes to their maturation (i.e., stronger colored colonies appear to mature the fastest merely due to their strong color and dramatic contrast to the background). The halftime from the superfolder assay would in that case be incorrect, a theory which might be supported by the odd behavior displayed in its scatter plots where all the green mutants never seemed to reach full maturation (Figure 6 and 7). All previous information we have used when claiming that the green mutant is slow-maturing is not from proper maturation assays (Liljeruhm et al. 2018, Sydney iGEM 2016), but rather observations of how fast bacteria become colored. The problem with this is that several other factors can affect this, such as the expression level of the protein, protein folding (immediate folding into functional shape vs. formation and refolding of inclusion bodies) and the extinction coefficient. In other words, even though amilCP green might develop color slowly it does not necessarily mean that it matures slowly.


• The analysis has a bias towards weaker colors (i.e., weaker colors, such as amilCP green, appear to mature the fastest). Since the other weak color, Gly157, also has a fast maturation halftime from this run, one possibility could be that there is some issue in the detection of smaller differences in grayscale values. This seems quite unlikely however, as the scatter plots look good and similar to other stronger colors, and the analysis is designed to only consider relative change in intensity. Further, the relatively weak colored pink Asp157 mutant has the slowest halftime according to this assay, while the strong colored mRFP1 has the fastest.


• There is another error somewhere, either in the preparation of the sample, the measurement or in the analysis. If we would have had more time, we would have redone this experiment to assess whether the half time of amilCP green is accurate or not.

Aside from the unexpected green amilCP halftime, the light purple Gly157 (NNN157 E) mutant seems to mature the fastest, faster than the wild type amilCP, while the purple Ala157 (NNN157 D) mutant is slightly slower. The pink Asp157 mutant is the slowest in this assay. The varying maturation half times between the mutants further indicates that position 157 not only impacts the color of amilCP, but also is important for the maturation rate of the protein.

Superfolder mutations

---

Sequencing confirmed that the introduction of mutations was successful for both the one point and the two point mutation (Figure 15). Amino acids 37-40 of amilCP were mutated from the original EQTV sequence to TQTV for SF1 and TQNV for SF2.

The results of the maturation assay indicates that no improvement of the maturation rate was made by the superfolder mutations (Figure 16). This result is in line with what was documented also for the blue amilCP, which was tested by Bao et al. (2020). However, we can not confidently draw this conclusion from the assay, since the scatterplots of the green amilCPs, both ordinary and superfolders, behave oddly and indicate that they did not fully mature during the 48 hour assay (Figure 6 and 7; Appendix 3). The scatterplots of the green amilCPs from the color changing run are widely different and show that by the same time the superfolder run ended, at hour 48, they were already fully matured (even several hours before that; Appendix 4). It is difficult to explain why we see such a large difference between the green amilCP between the runs. Further investigation into this would be of great interest. Until then, we can not confidently draw any conclusions about the impact of superfolder mutation on the maturation rate.

Random mutagenesis

---

Two rounds of error prone PCR were performed using two different sets of primers. In the first round, random mutations were introduced throughout the entire coding sequence of green amilCP and the upstream promoter region. Since unwanted mutations in the promoter region could be detrimental to the expression of the gene, another round of error prone PCR was performed where only the coding sequence was subjected to random mutagenesis.

Both rounds of error prone PCR gave promising results. Agarose gel analysis of the PCR products revealed bands of expected sizes of around 800 bp (Figure 17). Increasing the manganese concentration resulted in less PCR product, as can be seen by the decreasing intensity of the bands on the gel. This is expected since increasing levels of manganese reduces the performance of the polymerase until it no longer can bind to the template and fails to replicate the full-length DNA fragment. A manganese concentration of 0.4 mM was clearly too high, as almost no DNA was amplified.

Unfortunately, transformation of the PCR products into E.coli yielded no colonies at all and so the error prone mutants were never expressed and could not be analyzed further. Several transformation attempts were made, and some colonies did in fact grow on a few plates. However, colony PCR revealed that the plasmids had not been internalized by the cells, implying that it was only contamination growing on the plates (Figure 18).

The results from the transformation attempts suggest that the assembly into intact plasmids failed. Since the transformation of our positive control mRFP1 succeeded, there is little reason for us to believe that the transformation step is the problem. Instead, it is more likely that the error prone PCR product did not ligate with the backbone vector properly.

For digestion and ligation to work properly, it is critical that the enzyme stocks used are functional. To test the activity of the restriction enzymes, an experiment was conducted where mRFP1 plasmids were digested with different restriction enzyme combinations. Analysis of the digests on an agarose gel showed that the enzymes were able to cut the DNA, although not 100% effectively as bands corresponding to uncut DNA were still visible (figure not shown). Gel excision and purification was used to separate the digested DNA from the undigested. This worked well for purifying the digested backbone vector since linear and the circular/supercoiled plasmid DNA travel at different speeds through the agarose and produce distinct bands on the gel. However, it is much more difficult to purify the digested error prone PCR product, let alone to verify that the digestion even worked. Since the digested PCR product and the intact PCR product only differ by a few base pairs in size, they will not separate clearly on the gel. Therefore, we cannot be certain that the PCR product was digested properly. Moreover, we cannot rule out the possibility that the ligase might have been the problem. If we would have had more time in the laboratory, a ligase test experiment would have been a priority. In hindsight, all experiments should have begun with testing out the material to ensure that everything was working as intended.

Turbofolder mutations

---

In an effort to improve the maturation rate of green amilCP, programs predicting protein folding and simulating molecular dynamics were used to find potential mutations that could create a pore in the protein. The idea was to create a pore through which oxygen could easily access the chromophore. Through site-directed mutagenesis, the leucine (Leu) at position 195 was supposed to be substituted with four different amino acids: glycine (Gly), valine (Val), alanine (Ala) and isoleucine (Ile). These were the so-called “Turbo mutations”.

Unfortunately, due to some errors in the primer design and sequencing the effect of the Turbo mutations could not be assessed. If we would have had more time in the lab, we would have prioritized redoing these experiments.