Overview

BUCT mainly requires the synthesis of caffeine and 2-phenylethyl alcohol (2-PE) on the basis of fatty acid consumption. We successfully knocked out a gene that inhibits fatty acid oxidation in E. coli, thereby enhancing the free fatty acid consumption function of E. coli Nissle 1917. After the first-generation synthesis of 2-PE, we compared the yield with JLU-China and set the second-generation plan. The synthetic yield of caffeine is low, but we do see some accumulation of the precursor theobromine.

Based on the comprehensive consideration of the growth and effective yield of E. coli Nissle 1917, we also calculated and modeled the highest bacterial density that could inhibit Malassezia and satisfy the effective amount of product.

Fatty acid consumption

We have successfully knocked out the fadR in E. coli Nissle 1917 through CRISPR/Cas9.

The strain was also used for the synthesis of 2-phenylethyl alcohol and caffeine. According to the growth monitoring data of this strain in the fermentation process, we fitted a growth curve to calculate the yield of E. coli Nissle 1917 in the actual process.

Synthesis of 2-Phenylethyl Alcohol (2-PE)

Figure 1. 2-PE fermentation samples

The first round of 2-PE synthesis has been completed. The concentration of the product can be calculated by HPLC. E. coli Nissle 1917 was able to synthesize 2-PE successfully by adding glucose, which indicated that we successfully opened the synthesis path in Nissle, although the yield of BUCT in de novo synthesis was only about 60% of that in JLU-China, which also synthesized 2-PE. However, the small molecule synthesis of aromatic phenols, which we used as an example of 2-PE, can be used as a prerequisite for more flavor modifications in the future.

Figure 2. Standard curve for measurement of 2-PE by HPLC

The confidence of the 2-PE standard curve is very high, and we can directly filter the samples after fermentation without derivatization.

Figure 3. Yield of 2-PE fermentation in E. coli Nissle 1917

In the future, the synthesis of 2-PE will continue to be optimized, and we will try to use new genes and add phenylalanine to promote the synthesis of 2-PE.

We modeled aroG-fbr homology and aligned the molecular model of the enzyme with the original type:

Figure 4. BUCT homology modeling aroG-fbr, original type on the left is blue and BUCT mutant on the middle is pink.

The mechanism by which phenylalanine inhibits AroG activity is very complex. The new AroG binds to Phe and the whole molecule undergoes a dramatic conformational change. There is a study shows that our mutant aroG^fbr was shown to alleviate the inhibitory effect of the phenylalanine analogue 4-fluoro-DL-phenylalanine (p-EP). p-EP is a stronger inhibitor of aroG than Phe, and when aroG is inhibited by p-EP, E. coli cannot grow. We visualized this part of the conformational change by homology modeling. We hypothesized that the conformational change of the enzyme protein caused by the mutation might be the main reason for the reduced inhibitory effect of Phe. In the model comparison results, we found that the mutated molecular pocket appears to be covered by a larger lid.

Figure 5. The red and dark blue sections represent the molecular conformational changes after the aroG sequence mutation. The spatial position of residues in front of the molecular cavity increases, and the entrance of Phe as an inhibitor into the binding site is like blocked.

Synthesis of Caffeine

Figure 6 Caffeine fermentation sampling record

In order to better detect the synthesis of caffeine, the cumulative amount of theobromine was also measured when the fermentation products were measured, and the standard curves were made respectively, as shown in Figure 6 and Figure 7.

Figure 7. Standard curve for measurement of theobromine by HPLC

Figure 8. Standard curve for measurement of caffeine (CA) by HPLC

The above standard curves are in line with the standard and can be used for the conversion of theobromine and caffeine production. Finally, the maximum yield of theobromine was 7.16mg/L and the maximum yield of caffeine was 1.83mg/L. In the results, we can find theobromine was more productive than caffeine, and it was speculated that the lower methylase activity in the last step resulted in the accumulation of theobromine and affected the synthesis of caffeine. Whether the final output meets our requirements is calculated and forecast through modeling (see Yield prediction and promoter screening below).

Figure 9. Yield of theobromine fermentation in E. coli Nissle 1917

Figure 10. Yield of caffeine fermentation in E. coli Nissle 1917

Yield prediction and promoter screening

Figure 11. Strain growth fitting curve

In the modeling work, the strain growth fitting curve represents the accuracy of our calculation to a certain extent. And based on that, we predicted caffeine production.

Figure 12. Product formation kinetc

By using the Logistic and Luedeking-Piret equations, we have succeeded in constructing a model that can predict caffeine production, and the resulting model fits the experimental data well, proving the feasibility of the model.

It can be concluded that the engineered bacteria are able to produce enough caffeine in practice to exceed the 0.4 mg/L effective concentration threshold. We have also created a model of the strain competition mechanism and used it to assist the design of the suicide mechanism. We calculated that the triggering OD=0.21, at which the strain could suppress Malassezia at various concentrations effectively, while at the same time initiate mazF gene expression, induce apoptosis and ensure that the engineered bacteria do not spread excessively.

We invite you to the Model page.