Our project aims to engineer a coral-localized E. coli to produce and regulate biofilm, specifically through the mazE/F genes, the csg (or curli) operon (for biofilm production), and the metJ regulon(for biofilm regulation) respectively. Throughout our experimentation, we followed an iterative engineering cycle, the steps include: Design, Build, Test, and Learn. This cycle is especially apparent in our cloning experiments; in which we designed and assembled our desired plasmids.

The Csg construct was designed in benchling and the sequence was provided to IDT for synthesis. Due to its prolonged length, the construct was synthesized in two fragments with the separation occurring at 1028 bp, making fragment 1 (csg 1) 1027 base pairs and fragment 2 (csg 2) 2103 base pairs. We used the backbone pSB1C3 as its vector due to its high copy number and assembled the fragments using biobrick assembly.

First, we performed amplification PCR to amplify the fragments synthesized through IDT. We struggled to amplify the product, and even when we obtained the product, they were all of low concentration and high impurity.

To optimize our PCR condition, we tried different DNA polymerases, such as Phusion, Q5, and Primestar polymerases. We also altered the PCR cycles. Since we hypothesized that our construct might have higher GC content that might result in secondary structures formed during annealing, we added DMSO, which is commonly used to reduce secondary structures. DMSO can also help relax supercoiled plasmids by increasing single-stranded regions, which allows enhanced topoisomerase I activity.

Despite such changes, our colony PCR results continued to be negative, showing band lengths much lower than the expected 5177bp (Figure 1).

We attempted to knock out the csg genes from the E. coli genome using homologous recombination to increase the accuracy of our assays involving biofilm. Because E. coli naturally produces biofilm, when we express our Csg construct in E. coli, we cannot decipher whether the biofilm is produced naturally by E. coli or our Csg construct, nor quantify the potency in producing biofilm by our synthetic Csg construct.

To test our Csg construct that produces biofilm, before we transform our Csg construct into E. coli, we need to knock out original biofilm genes from the E. coli genome to increase the accuracy of our testing.

We attempted to knock out the csg operon, responsible for producing and secreting biofilm, from the E. coli genome using homologous recombination ^[1].

The homologous recombination process involves:

1. Template plasmids (pKD13) carrying Kanamycin-resistance genes flanked by FRT (FLP recognition target) sites.
2. We generated PCR products by using primers with 36- to 50-nt extensions that are homologous to regions adjacent to the csg operon.
3. By electroporating the PCR product into E. coli strain MG1655 containing the helper plasmid pKD46 encoding the Red recombinase, the KmR gene swaps with the csg operon on the E. coli genome.
4. KmR was then eliminated by using the helper plasmid pCP20 encoding the FLP recombinase.

We obtained the materials used for KO homologous recombination from Dr. Masayuki Hashimoto of the Institute of Molecular Medicine at the National Cheng Kung University:

1. E. coli strain MG1655 contains pKD46 plasmid that aids in replacing the csg operon with our amplified Kanamycin-resistant gene.
2. pKD20 plasmid encoding the FLP recombinase to eliminate the Kanamycin-resistant gene.

Due to time constraints from delays to our cloning process, we did not complete the designed gene knockout process.

The MetJ construct was designed in benchling and the sequence was provided to IDT for synthesis. Due to its prolonged length, the construct was synthesized in 5 fragments with the separations occurring at 1703 bp, 2441 bp, 3540 bp, 4301 bp, and 5054 bp.

We used pET28a as our plasmid because of its low copy number, to prevent an over-expressing of MetJ proteins that may plummet biofilm production, too quickly and also because we plan on co-transforming both the plasmids for the Csg construct and MetJ construct, so we needed a plasmid with a different antibiotic resistance gene from pSB1C3 for screening.

First, we performed amplification PCR to amplify the fragments synthesized through IDT. We struggled to amplify the product, and even when we obtained the product, they were all of low concentration and high impurity.

To optimize our PCR condition, we tried different DNA polymerases, such as Phusion, Q5, and Primestar polymerases. We also altered the PCR cycles. Since we hypothesized that our construct might have higher GC content that might result in secondary structures formed during annealing, we added DMSO, which is commonly used to reduce secondary structures. DMSO can also help relax supercoiled plasmids by increasing single-stranded regions, which allows enhanced topoisomerase I activity.

Despite such changes, we continued to obtain PCR products of low concentration and purity. We attempted to perform Gibson assembly to assemble the fragments into the vector pET28a, but when we ran Colony PCR to confirm the fragment size, we did not see anything on the gel. We hypothesized that the inability to obtain product from Gibson assembly is due to the high amount of fragments that may decrease the efficiency of Gibson assembly, so we ran overlapping PCR to connect our fragments in pairs (1, 2+3, 4+5) before performing Gibson Assembly. We failed to obtain any product (Figure 5), even after changing our PCR conditions: different # of PCR cycles, different polymerase, and different amounts of DNA product.

Because of the repeated failures in running amplification PCR and overlapping PCR, even after altering our experimental conditions (PCR conditions), we decided to redesign our primers. Our main issue with our primers is that our overlapping primers had too short of a homologous region (20bp), which made our primers anneal poorly. We redesigned our primers by increasing the homologous regions of our overlapping primers to 30 bp and decreasing the GC content in the primers to further prevent non-specific binding and formation of secondary structures during amplification PCR (Table 2).

After our new primers arrived in late August, we successfully re-amplified our fragments (Figure 6). Our amplification PCR products’ concentrations improved significantly (Table 2).

We then performed Gibson assembly to ligate the 5 fragments and the vector (pET21A), which we then transformed into E. coli DH5α. We failed to obtain the correct size during colony PCR to confirm the fragment size.

Due to time concerns, we decided to clone the fragment that contains the metJ gene (fragment 3) synthesized by IDT and into the pJET vector, and transform it into E. coli strain DH5α. This allows us to at least test if the MetJ protein can successfully repress biofilm production. We successfully cloned the MetJ fragment 3 into the pJET vector, and performed colony PCR to confirm the fragment size (Figure 7). The sequencing results showed that our construct was successfully cloned.

To increase the efficiency of amplification PCR and Gibson assembly, we will review our fragments to possibly lower the GC content at the ends.

The Biosafety construct was designed in benchling and the sequence was provided to IDT for synthesis. Due to its prolonged length, the construct contains 4 fragments with the separations occurring at 775 bp, 2130 bp, and 2802 bp. We used the backbone pET21a as a vector. We attempted to assemble the fragments using Gibson assembly. We were able to successfully amplify all but the third fragment (Figure 8), even after redesigning our primers. Due to time concerns, we will further complete this construct in the future.

To improve cloning efficiency for future cloning success, we will redesign our fragments to lower the GC content at the ends, increasing the efficiency of amplification PCR and Gibson assembly.

[1] Datsenko, K. A., & Wanner, B. L. (2000). One-step inactivation of chromosomal genes in Escherichia coli K-12 using PCR products. Proceedings of the National Academy of Sciences of the United States of America, 97(12), 6640–6645. https://doi.org/10.1073/pnas.120163297