Cycle 2: Promoter Prediction

We developed a computational tool (Prokaryotic Motif-Based Promoter Strength Comparison Analysis Program, PPAP) to predict the strength of 13 bacterial promoters in Lactococcus lactis based on conserved motifs. Our goal was to identify key motifs, specifically the -10 (TATA box) and -35 (Sigma factor binding site) sequences, which play a crucial role in gene transcription. By analyzing GC content and motif placement, we created a custom scoring system to rank promoters based on predicted efficiency. This computational approach was intended to assist in selecting strong promoters for gene expression, minimizing the need for extensive lab work and experimental iterations.

• Design: We aimed to develop a computational tool to predict the strength of 13 bacterial promoters in Lactococcus lactis based on conserved motifs. Our goal was to identify key motifs, like the -10 and -35 sequences, which play a crucial role in gene transcription. By analyzing GC content and motif placement, we created a custom scoring system to rank promoters based on predicted efficiency. This computational approach was intended to assist in selecting strong promoters for gene expression, minimizing the need for extensive lab work and experimental iterations.

• Build:  We implemented the analysis by running our script on 13 selected promoters. Each promoter’s GC content and the presence of conserved motifs were calculated, and these features were integrated into a scoring function. Promoters were then ranked by strength, with nisA_op scoring the highest at 13.28. A promoter with a higher custom score should, theoretically, be more effective in initiating transcription and thus be a stronger promoter.

• Test:  To validate our computational predictions, we plan to test these Lactococcus lactis promoters in vivo using BCA Assay and enzyme-linked immunosorbent assay (ELISA). Most of the promoter strength results were in line with our predictions with the absolute p-value lower than 0.1, despite two exceptions observed in nisF and P32 promoters (with p-values of 0.166256 and 0.10916 respectively), concretely validate the promoter prediction model.

• Learn: Through this project, we have learned that bioinformatics tools can significantly streamline the process of promoter selection. Our custom scoring system offers a computational approach to predict promoter strength based on conserved motifs and GC content. (Results for prediction: nisA_op: 13.28 P48: 12.50 P11: 11.00 cc: 9.39 nisR: 8.96 nisR_op: 8.94 P5: 8.77 P3: 8.73 nisA: 8.32 P8: 8.28 nisF: 6.36 PTCIIC-celA: 0.36 P32: 0.00)

Cycle 3: NisRK Overexpression

Overexpression of nisRK with plasmid vectors can promote PnisA activity by utilizing the self-regulation of nisin in NICE system, thereby increasing nisin yield. This scheme not only can verify the feasibility of our genome editing scheme which aims to edit the nisR promoter, but also can be used as a simple and independent way to increase the expression of nisin.

• Design： The membrane-located histidine kinase NisK senses the inducing signal nisin and autophosphorylates, then transfers phosphorous group to intracellular response regulator protein NisR which activates nisA promoter to express the downstream gene.

• Build:  We added Myc-tagged nisR CDS and HA-tagged nisK CDS into the pMG36e plasmid with PnisA-GFP, aiming to detect the effect of nisRK on nisA promoter activity through changes in GFP expression level, and verifying nisRK expression through detection of respective tags.

• Test:  We aim to detect expression of both AcGFP reporter gene and nisK through Western blot, targeting Flag tag and HA tag, respectively. As a semi-quantitative method, Western Blot can verify the expression of nisRK regulation system while also providing some clues on the changes in nisA promoter activity in the presence of nisRK.

• Learn:  The experimental results show that nisRK was successfully expressed. In addition, the increase in GFP production caused by overexpression of nisRK was detected, which can indicate possible changes in nisin production guided by the same promoter, verifying the feasibility to increase nisin production by increasing nisRK expression through genome editing. Also, this solution alone can also be used as a simple method to increase nisin yield.

Cycle 4: Genome Editing

In Genomic editing engineering cycle, we discuss the design of approach to improve nisin production. In the design phase, original promoters are designed to be replaced with strong constitutive promoters, utilizing CRISPR-Cas9 for precise genome editing. The build phase focuses on constructing an appropriate plasmid (pMG-Cas) for transformation into Lactococcus lactis. Testing evaluates nisin yield via qPCR and protein assays, with GFP reporters to assess promoter activity.

• Design: Our goal is to enhance nisin production in Lactococcus lactis by replacing weak or inducible promoters in the nisin gene cluster with strong constitutive promoters. Specifically, PnisR, the promoter responsible for expression of the nisRK regulatory pathway, will be substituted with a stronger constitutive promoter to increase transcriptional activity and thus nisin yield. Homology arms for genome editing was obtained based on whole genome sequence of the Lactococcus lactis strain ATCC11454. The genome editing will be performed using a Red/ET assisted CRISPR-Cas9 system. The plasmid pMG-Cas, which incorporates pMG36e’s pWV01 ori and Ery resistance gene into the CRISPR/Cas9 plasmid pCas, was design be used as the CRISPR delivery vehicle, ensuring stability and compatibility with the host system.

• Build: The plasmid pMG-Cas was be constructed by replacing the pUC ori and Kan resistance gene from pCas with the pWV01 ori and Ery resistance gene from pMG36e. This allows the plasmid to function effectively in Lactococcus lactis. The homology arms flanked the strong constitutive promoter, and a Cas9-sgRNA system was designed to target PnisR for substitution. Following plasmid assembly, Lactococcus lactis cells would be transformed with the pMG-Cas plasmid. The cells would be cultured at 30°C to maintain plasmid stability. PCR amplification can then confirm successful integration of the promoter at the desired locus, followed by sequencing to ensure the accuracy of the modification.

• Test:  Whether the knock-in was successful will be determined by PCR after the transformed plasmid is lost. Nisin production levels was designed to be measured post-transformation to assess the effectiveness of the promoter substitution. A secondary experiment would utilize GFP as a reporter to monitor the activity of the new constitutive promoter in real-time, providing an indirect measure of its strength compared to the original inducible promoter. This would help identify any discrepancies in promoter activity or expression levels.

• Learn:  Data from PCR and nisin production assays would be analyzed to determine whether the new promoter was successfully knocked in and whether it significantly enhances expression. If successful, the promoter substitution can be considered validated. Possible bottlenecks still need to be identified such as low CRISPR efficiency or plasmid instability, which influence the selection of promoter or the construction of plasmids.

H2O2 Solution

One team is working on enhancing the hydrogen peroxide production of L. johnsonii to inhibit S. aureus, which is linked to toxic shock syndrome (TSS). We aim to engineer L. johnsonii by overexpressing two genes, LJ_0548 and LJ_0549, which encode subunits of FMN reductase (FRedA/B). These genes were found to significantly impact H2O2 production, with overexpression increasing it. The research was done through the following DBTL (Design, Build, Test, Learn) Cycles. Using the pMG36E plasmid, we cloned the genes under the control of a strong P32 promoter and transformed the plasmid into L. johnsonii (ATCC 33200). To test expression, we tagged LJ_0548 with Flag and LJ_0549 with 6xHis Tag. The engineered strain could potentially reduce TSS risk when introduced into tampon products.

Cycle 5: H2O2 Enhancement

• Design: pMG36E is a lactic acid bacteria expression vector that contains the P32 promoter, an erythromycin resistance gene and a multiple cloning site. (van de Guchte, Van der Vossen, Kok and Venema) We designed our plasmid based on this backbone.

The plasmid is designed to overexpress a protein dimer FMN reductase, which consists of FRedA and FRedB, thus increasing the production of hydrogen peroxide of L. johnsonii.

• Build: We introduced genes encoding the proteins FredA and FredB into the backbone pMG36E. Both coding sequence were optimized using Genscript’s codon optimization service Gensmart. To detect facilitate testing detect the expression of these two genes, we attached Flag tag and 6xHis tag respectively.

Fig. 1 pMG36E with gene LJ_0548 and LJ_0549.

• Test:  

Fig.  2 H2O2 expression test and protein expression level test of L. johnsonii.

a) Description of the four different groups about their strain chosen and culturing environment. pMG36E stands for the strain with the plasmid backbone and pMG36E_LJ0548-0549 stands for the overexpression strain.

b) H2O2 production measured using phenol red assay for four groups.

c) Western Blot result for Flag Tag (up) and 6xHis tag (down). For each group, 2 replicates were done.

d) Result of ELISA test for Flag tag which is attached to the gene LJ_0548 after culturing for 16h. Group B2 was ignored because of too low protein concentration.

Four groups with different strains and done under different conditions were examined for H2O2 production and gene expression level. (Fig. 2a) The overexpression of the two genes which encode the FMN protein dimer did not show significant difference in aerobic groups, but showed a notable increases in H2O2 production in anaerobic groups after growth of 12 hours. (Fig. 2b)

For the Western Blot result of gene LJ_0548, no significant differences were observed for anaerobic groups and the result was not trustworthy due to too much unspecific binding of proteins whose length did not march our target (46.29kD). (Fig. 2c) Under aerobic conditions, the expression levels of the protein FRedB encoded by gene LJ_0549 were moderately increased. Although this result means gene LJ_0549 in the plasmid successfully expressed, it is not significantly elevated.

To validate protein expression, ELISA was carried out and the outcome revealed a significant downregulation of protein levels in the insertion groups under both conditions, aligning with the WB observations for the 6xHis tag. (Fig. 2d)

• Learn:  Our data suggest that anaerobic conditions favor higher H2O2 production, and the overexpression of LJ_0548 and LJ_0549 amplifies this effect. Aerobic conditions, on the other hand, limit H2O2 production, regardless of gene overexpression. The insertion of the two genes leads to a notable increase in H2O2 production in anaerobic environment after certain period of growth time.

Upon analyzing the relationship between H2O2 production and protein expression, we observed a negative correlation. While H2O2 levels increased following gene insertion, the corresponding protein levels for LJ_0548 and LJ_0549 decreased. We hypothesize that this reduction in protein levels may be due to the consumption or degradation of these enzymes after H2O2 production. Additionally, this may represent a regulatory mechanism by L. johnsonii to prevent excessive H2O2 accumulation, protecting the bacteria and its environment from oxidative damage.

Cycle 6: Protein Docking Analysis

• Design: FMN (flavin mononucleotide) reductase is a protein dimer which is made up of the 2 proteins we inserted (nfr1/2, or FRedA/B). Because the function of the two protein is still not fully understood and research showed that the loss of the two proteins would disable the bacteria’s ability to produce hydrogen peroxide, we use protein function prediction tool and protein docking online services to predict the function of this protein dimer. We hoped that the result could give us some insight on the mechanism of hydrogen peroxide production in Lactobacillus johnsonii.

• Test: First, we uploaded our 2 protein sequences (nfr1/2) to STRING, a database of protein-protein interactions.

In our current organism L. johnsonii, 9 predicted functional partners of nfr1 and nfr2 were chosen for analysis (Fig. 3).

Fig. 3 Nine predicted functional partners for our reductase which were found by STRING. AOG26301.1 is nfr1.

ClusPro, a web-based tool was used for protein docking for conducting our analysis of the 9 predicted functional partners. The 9 partners which were found by STRING were performed protein docking with the FMN reductase.

For the FMN complex and Oxidoreductase 1 of L. johnsonii (BBP16_00775), the protein docking result shows that the surface representations of them fit together well, indicating good structural complementarity. (Fig. 4) The highlighted residues (red and yellow) indicate key interaction sites and interacting surfaces, contributing to the stability and specificity of the interaction.

Fig. 4 The docking result between the FMN complex (in cyan) and the Oxidoreductase 1 (in magenta).

Also, ClusPro clusters the docking results to identify similar binding modes and provides scores for these clusters. (Table. 1) Among all the partners we found in Fig. 3, this cell surface protein shows the lowest energy. This suggests that the docking model is the most stable among the 9 partners and biologically relevant.

Table. 1 The scores provided by ClusPro for the protein docking result for FMN complex and Oxidoreductase 1.

• Learn:  The docking result between FMN complex and the Oxidoreductase 1 of L. johnsonii (BBP16_00775) shows significant surface contact. FMN may play a role in catalyzing the transfer from reductants to oxidant and work with oxidoreductase inside L. johnsonii to deal with the problem of hydrogen peroxide accumulation. The structural complementarity observed suggests that the docking model is plausible and may provide hint for the analysis of metabolism chain.