We used homologous recombination to construct the pCDFDuet-tktA-aroF-aroE plasmid. First, using PCR technology, we designed corresponding primers and amplified the target fragments of tktA, aroF, and aroE using Escherichia coli K12 as a template. We also amplified the pCDFDuet plasmid backbone using a laboratory-preserved plasmid. As we can see from Figure 1A, the agarose gel electrophoresis results after amplification showed that the amplified lengths of the backbone and target fragments were all in line with the expected sizes. The pCDFDuet plasmid backbone is 3781 bp in length, tktA is 1995 bp in length, aroF is 1071 bp in length, and aroE is 819 bp in length.

After obtaining the empty vector and fragments of corresponding lengths, we recovered the agarose gel to obtain the purified vector and fragments, and obtained the concentration of these gene fragments through the nanodrop instrument. After obtaining the corresponding concentration, we used homologous recombination to allow the vector and fragments to be recombined together. Then we transformed the recombinant product into the E. coli cloning strain TOP10. As shown in Figure 1B, after the transformation and overnight growth, we saw that there were single colonies with successful transformation. However, whether the transformation was successful or not, we need to perform corresponding verification.

First, by colony PCR, we randomly selected 4 independent single colonies of the same size on the plate and used colony PCR primers for identification. The length was designed to be 2102bp. In PCR After amplification and identification by agarose gel electrophoresis, we found that the size of the amplified fragments met our expectations. Therefore, these single clones were considered to have successfully transformed with the recombinant plasmid.

Finally, we need to expand and culture the corresponding correct single clones, extract the plasmids, and send them to a biological company for sequencing. The purpose of this step is to ensure the position and order of our vector and connector, and whether there are any base deletions or mutations. Figure 1D shows that our construction results are correct. At this point, our first recombinant plasmid was successfully constructed.

Figure 1. A. Electrophoresis results of plasmid backbone and target fragment, B. Recombinant plasmid transformed into cloning strain TOP10, C. Electrophoresis results after single clone colony PCR, D. Plasmid sequencing results

We used E. coli K12 as a template to amplify other target fragments. First, we designed corresponding primers and then used cultured E. coli K12 as a template to amplify the target genes TyrA, hpaB, and hpaC. The plasmid backbone was also obtained by inverse PCR. After the amplification reaction, the amplified products were identified by agarose gel electrophoresis. The results showed that the plasmid backbone and the three target genes were successfully amplified. The backbone length was 5420 bp, the length of TyrA was 1122 bp, the length of hpaB was 1563 bp, and the length of hpaC was 513 bp, as can be seen in Figure 2.

Figure 2. Electrophoresis results of pETDuet and TyrA(K12), hpaB and hpaC

After obtaining the gene fragment of the correct size, we need to recover the purified vector and fragment by gel recovery. Then, we use homologous recombination to construct the complete recombinant plasmid. Next, we need to transform it into a cloning strain and, after overnight growth, select the grown single clones for colony PCR identification. As shown in Figure 3, we selected six single clones and amplified them using the designed primers. The results showed that the amplified length was 1500bp, which is consistent with the expected length, indicating that the transformation process was successful.

Figure 3. Results of monoclonal plates and colony PCR after transformation

Finally, we inoculated the successful transformants into seed solution, cultured them, extracted the plasmids, and sent them to a biotechnology company for sequencing. The results are shown in Figure 4. The sequencing results show that the constructed recombinant plasmids have no base mutations or deletions.

Figure 4. Plasmid sequencing results

Next, we constructed the pETDuet-TyrA(SO)-hpaB-hpaC plasmid. Compared with pETDuet-TyrA(K12)-hpaB-hpaC, the difference between the two plasmids is the source of TyrA. Therefore, we used homologous recombination to replace TyrA based on pETDuet-TyrA(K12)-hpaB-hpaC. We designed two pairs of primers, one to amplify the TyrA gene from the new source, and the other pair of primers to amplify the rest of the pETDuet-TyrA(K12)-hpaB-hpaC except TyrA(K12). After amplification, agarose gel electrophoresis was performed for identification, as shown in Figure 5A. As shown in Figure 5B , the plasmid backbone and TyrA(SO) amplification results after electrophoresis were consistent with the expected lengths: the plasmid backbone was 7535 bp long, and TyrA(SO) was 1152 bp long. The two fragments were then recovered and purified from the gel. After homologous recombination, the recombinant products were transformed into E. coli TOP10. As shown in Figure 5B , single clones emerged after overnight growth. To verify the correctness of the transformed single clones, we performed colony PCR identification. Twelve single clones were selected for identification. In Figure 5C , the colony PCR result showed a size of 2100 bp, demonstrating that the transformed single clones all contained the correct plasmid. The correct single clones were then inoculated into seed solution, and the plasmids were extracted and sent to a biotechnology company for sequencing. Sequencing results confirmed that the constructed plasmid was successful.

Figure 5 A. Electrophoresis results of plasmid backbone and target gene fragments, B. Plate of recombinant plasmid transformed into E. coli, C. Electrophoresis results of monoclonal colony after PCR, D. Sequencing results of recombinant plasmid

When constructing pETDuet-TyrA(HI)-hpaB-hpaC, we used a similar approach to pETDuet-TyrA(SO)-hpaB-hpaC. We first designed primers to amplify the TyrA(HI) target gene, and another pair of primers to amplify the backbone. After amplification, we verified the success of the amplification by agarose gel electrophoresis. As shown in Figure 6A, the amplified vector and target gene lengths were consistent with the expected values, with the backbone being 7535 bp and the target gene being 1146 bp. Following amplification, the two were ligated by homologous recombination and transformed into the corresponding E. coli cloning strain. After overnight growth, as shown in Figure 6B, 12 single colonies were selected for colony PCR analysis to confirm the success of the transformants.

Figure 6 A. Plasmid backbone and target gene electrophoresis results. B.plate image of the transformed single colonies.

After selecting the corresponding single clone for colony PCR, we performed electrophoresis analysis on the amplified product. Figure 7A shows that the colony PCR result met the expected length of 2000 bp. Furthermore, we selected a single clone that successfully obtained colony PCR and inoculated it into seed solution. The plasmid was then extracted and sent to a biotechnology company for sequencing. The results are shown in Figure 7B. The sequencing results indicate that the recombinant plasmid construction was successful, with no base mutations or deletions.

Figure 7 A. Plasmid backbone and target gene electrophoresis results. B.plate image of the transformed single colonies.

After constructing Plasmid 1 and the three TyrA plasmids 2, we need to transform two plasmids into a new E. coli expression strain. Because the production of L-DOPA requires the simultaneous function of all six genes, we first transform two plasmids into E. coli BL21 (DE3), as shown in Figure 8. After overnight growth, we see that the double-transformed host bacteria have corresponding monoclonal colonies growing. We will then perform colony PCR to verify whether the double plasmid transformation is successful.

Figure 8. The corresponding single colony plate image after plasmid 1 and plasmid 2 were transformed into BL21 (DE3) together

After the double plasmid transformation is completed, we need to perform colony PCR identification. Since there are two plasmids, two pairs of primers need to be designed to identify each plasmid separately. Therefore, colony PCR is performed on each single clone of strain1-3. The corresponding lengths are: plasmid 1 of strain1 is 1000bp, plasmid 2 is 800bp, plasmid 1 of strain2 is 800bp, plasmid 2 is 1000bp, and plasmid 1 of strain3 is 1000bp, plasmid 2 is 1000bp. After the amplification is completed, we perform agarose gel electrophoresis. The results are shown in Figure 9. The lengths of all colony PCRs are in line with the expected size. Therefore, the double plasmid transformation is considered successful.

Figure 9. Electrophoresis of colony PCR results of three strains after double plasmid transformation

Before conducting the first round of fermentation tests, we needed to perform protein expression tests on the two plasmids to initially verify the normal expression of our six genes in E. coli. Therefore, we transformed BL21 (DE3) host cells with plasmid 1 (pCDFDuet-tktA-aroF-aroE) and plasmid 2 (pETDuet-TyrA(K12)-hpaB-hpaC) for protein expression. This round of protein expression testing employed rapid induction, using a temperature of 37°C, a rotational speed of 220 rpm, an inducer concentration of 1 mM IPTG, and a duration of 6 hours.

After induction, we collected the protein and divided it into the supernatant and precipitate of crude protein for SDS-PAGE identification to observe whether the expression of the target gene was successful. As shown in Figure 10, wells 1-4 are the supernatant of plasmid 1, 5 is the precipitate of plasmid 1, wells 6-8 are the supernatant of plasmid 2, and 9-10 are the precipitate of plasmid 2. Plasmid 2 contains inclusion bodies, and the proteins corresponding to the target genes are all successfully expressed. Although the target protein corresponding to plasmid 2 has inclusion bodies during the collection process, this may be related to the operational steps in the protein collection process. In addition, since we are strengthening the L-dopa biosynthesis pathway in the cell, we confirm that the expression has achieved our goal.

Figure 10. Target protein expression test using plasmids 1 (pCDFDuet-tktA-aroF-aroE) and 2 (pETDuet-TyrA(K12)-hpaB-hpaC). Wells 1-4 contain the supernatant from plasmid 1, and well 5 contains the precipitate from plasmid 1. Wells 6-8 contain the supernatant from plasmid 2, and wells 9-10 contain the precipitate from plasmid 2.

After confirming that our target proteins are expressed normally, we need to perform fermentation tests on the previously constructed strains carrying TyrA from three sources in order to obtain the optimal TyrA. The fermentation method is consistent with the protein expression conditions, and the fermentation time is 48 hours. Three replicates are set for each strain fermentation, as shown in Figure 11A. After the fermentation is completed, we use the L-DOPA Elisa detection kit for measurement. The sample preparation, reaction process, and reaction termination are shown in Figures 11BCD. Finally, we use a microplate reader to measure the absorbance value at 450nm Figures 11E.

Figure 11. A. Fermentation test of strain 123 (three replicates), B. Sample preparation, C. Reaction, D. End of reaction, E. Absorbance test

We first developed a standard curve, as shown in Figure 12A. We then calculated the L-DOPA content in the sample based on the standard curve and absorbance values. It can be seen that strain 1 had the highest L-DOPA production, as shown in Figure 12B. That is, TyrA derived from K12 showed better enzyme activity in the modified strain. For this modified strain, genes derived from similar species have better adaptability, so this gene was selected for subsequent RBS screening.

Figure 12. A. Standard curve of the ELISA test kit; B. L-DOPA production by three strains, where Strain 1 contains TyrA from K12, Strain 2 contains TyrA from SO, and Strain 3 contains TyrA from HI.

After confirming that TyrA from K12 was the most effective, we needed to further optimize the RBS sequence of this fragment. Since this gene is a key enzyme in L-dopa synthesis, we only replaced the RBS at the front end of this fragment, from the default RBS1: AAGGAG to RBS2: ATTAAAGAGGAGAA and RBS3: TCACACAGG, as shown in Figure 13.

Figure 13. Schematic diagram for replacing the RBS

To replace the RBS at the beginning of the TyrA gene in pETDuet-TyrA(K12)-hpaB-hpaC, we used inverse PCR. We designed the two RBS sequences to be replaced into two pairs of primers and then amplified them using the original pETDuet-TyrA(K12)-hpaB-hpaC plasmid as a template. Compared to the original plasmid, only the RBS sequence changed in the amplified plasmid. Since the length of the newly amplified plasmid was almost the same as the original, we did not perform agarose gel electrophoresis. Instead, we transformed the plasmid after amplification and extracted it for sequencing. The sequencing results in Figure 14 demonstrate that the new RBS sequence has been successfully replaced. The sequence on the left shows the replacement of RBS2, while the sequence on the right shows the replacement of RBS3.

Figure 14. Sequencing results after replacing the RBS sequence

After the RBS sequence replacement was completed, fermentation testing was repeated to determine which RBS sequence performed best. Three strains were inoculated into LBG (20 g/L glucose) medium at an initial OD of 0.05. After fermentation, product testing was performed to identify the optimal RBS sequence for the TyrA gene and improve expression of the key enzyme.

The fermentation testing steps were similar to those described above. Three replicates of different RBS strains were tested, as shown in Figure 15A. The assays were performed using a kit, with the reaction progressing and reaction completion times (Figure 15BC). Finally, the fermentation yields were calculated using a standard curve.

Figure 15. A Fermentation test of different RBS strains (three replicates), BC is divided into reaction progress and reaction completion

By substituting the absorbance value into the standard curve for calculation, we can obtain the amount of L-dopa produced by the strain under different RBS sequences. According to the bar graph in Figure 16, the three different RBS sequences do have a certain effect on gene expression. Among the RBS sequences we screened, the optimal expression level increased the strain's yield by 16.4%, reaching 500.3 mg/L.

Figure 16. The yield of levodopa corresponding to different RBS sequences

At this point, we have confirmed the optimal TyrA source (K12) and RBS sequence TCACACAGG. We then initially tested the effect of different glucose concentrations on L-DOPA production by fermentation in the BL21 (DE3) host strain. We set initial glucose concentrations of 10 g/L, 20 g/L, 30 g/L, 40 g/L, and 50 g/L, and the fermentation conditions were the same as before, After this, we will add some new parameters to explore the factors affecting levodopa production, such as inducer concentration and time.

Figure 17 shows that the yield of the strain increased with increasing glucose concentration, reaching a peak of 689.3 mg/L. However, as the glucose concentration continued to increase to 50 g/L, the yield began to decline. We speculate that the high glucose concentration affected the strain's metabolism, resulting in a decrease in yield.

Figure 17. Analysis of results at different initial sugar concentrations (optimal RBS sequence strain)

To further explore the factors affecting levodopa, we simultaneously investigated the inducer concentration, glucose concentration, and fermentation time. The glucose concentration was still set at 10 g/L to 50 g/L, with 5 gradients, and the inducer concentration was 0, 0.2, 0.4, 0.6, 0.8, 1.0 mM, with 6 gradients, and the fermentation time was 36 h and 48 h. The strain was inoculated into LBG medium with an initial OD of 0.05. After the fermentation was completed, the product was tested.

Figure 18. Effects of inducer concentration and glucose concentration on levodopa production at different times

Figure 18 shows the effects of inducer concentration and glucose concentration on levodopa production at different times. The overall trend is that when the inducer concentration is 0.8mM and the glucose concentration is 40g/L, the levodopa production is higher and significantly higher. When the inducer concentration is 0.2mM, the production is higher without adding the inducer concentration than after adding it. This may be due to contamination during fermentation. In addition, by comparing the overall situation at 36h and 48h, we can see that the levodopa production at 48h is higher.

The heat map in Figure 19 is actually a correlation graph. The response surface also tells us a trend. Finding the optimal inducer concentration and glucose concentration at different time points is consistent with the conclusion of the above bar graph. We can see that the optimal glucose concentration is 40g/L, while the inducer concentration is 0.8mM.

Figure 19. Optimal inducer concentration and glucose concentration at different time points

Finally, we analyzed the inducer concentration, glucose concentration and time together. In the three-dimensional results of Figure 20, we can see that 0.8-1.0mM, 40g/L, 36-48h, the fermentation results are better. Finally, we chose the conditions of 0.8mM, 40g/L, 48h for the fermentation of E. coli ECN.

Figure 20. Effects of different inducer concentrations, glucose concentrations, and time on levodopa

After we have explored the optimal inducer concentration, glucose concentration and time for L-dopa production in E. coli BL21 (DE3), the next step is to transform plasmid 1 and the optimized plasmid 2 into E. coli ECN, because our ultimate goal is to let probiotics produce L-dopa.

As shown in Figure 21, the transformation diagram on the left proves that plasmid 1 and plasmid 2 were successfully transformed into ECN. In addition, in order to further verify that there is no problem with the transformed plasmid, we used colony PCR for verification. The final amplified band length is 800bp, indicating that the transformation is successful. The next step is to select the correct single clone, expand the culture, and then induce fermentation to produce levodopa.

Figure 21. Plasmid 1 and plasmid 2 were transformed into E. coli ECN (left), and the electrophoresis results of colony PCR identification were shown (right).

We then fermented the successfully transformed ECN using the optimal conditions obtained in the BL21(DE3) host strain. After fermentation, we analyzed the fermentation samples using HPLC in triplicate. Figure 22 shows the peaks of the three replicates, revealing the concentrations of 1.064 g/L, 1.065 g/L, and 1.069 g/L, respectively.

Figure 22. Peak diagram of levodopa production after HPLC detection of ECN fermentation