. Engineering Success .
Overview
This year, XMU-China focused on the problem of soil frost and soil drought, endeavored to tackle and had solved problems during our engineering process, generating expected results.
Countless efforts were made to ensure our design can function to attain our goals. Among them, the two sections “obtaining a functional KpSP-GFP protein” and “harnessing pEcCas/pEcgRNA system to eliminate the cryptic plasmids of E. coli Nissle 1917” are typically worth mentioning. All the related data are recorded below. Hopefully, they will make some contribution to the iGEM community.
Section 1: Obtaining a functional KpSP-GFP protein
Introduction
The antifreeze proteins (AFPs) are utilized to solve frost in our design. Since AFPs function directly on the surface of the ice, whatever thermal hysteresis (TH) or ice recrystallization inhibition (IRI), it is necessary that AFPs should be secreted from the cytoplasm of our engineered bacteria to the extracellular circumstance. Hence, we hope to realize it by using a signal peptide from Kocuria sp.3-3, or KpSP, which directs target proteins to the extracellular circumstance.
Cycle 1: Using KpSP-GFP to test the secretion function of KpSP
Design
KpSP, found from Kocuria sp.3-3, has shown remarkable function to secret target proteins to the extracellular circumstance through the Sec pathway in a previous work (1). To firstly verify the function of KpSP, we fused it directly with GFP and then obtained KpSP-GFP to see whether GFP was secreted or not.
Build
There were four parts assembled into the vector: promoter (BBa_I0500), RBS (BBa_B0034), KpSP-GFP coding sequence, and terminator (BBa_B0015). We got the constructed plasmid successfully and transformed it into E. coli BL21(DE3), then the positive transformants were selected through colony PCR (Fig. 1) and sequencing.
Test
The fluorescence intensity of the bacterial culture and supernatant was measured to test whether KpSP can function well to lead to the secretion of GFP or not. In this case, BBa_I0500 (araC/pBAD, no gene downstream) was set as the negative control, while BBa_K4907100 (for expressing GFP intracellularly) which has no signal peptide's coding sequence was also tested to compare the fluorescence intensity. After being induced by L-arabinose for 6 hours, the fluorescence intensity of the three groups was measured.
Compared to the group expressing GFP intracellularly, no fluorescence was observed in the KpSP-GFP group, and even the ratio of fluorescence intensity to OD600 of the KpSP-GFP group showed no differences to the negative control group (Fig. 2), which indicated the fusion protein KpSP-GFP cannot work normally.
Learn
The result above showed that the fusion of KpSP might have negative effects on the folding or maturation of GFP, both of which are vital for fluorescent proteins to exhibit fluorescence normally (2).
Cycle 2: Using KpSP-GFP added a linker to test the secretion function of KpSP
Design and Build
Based on the result of Cycle 1, we added a linker between KpSP and GFP (BBa_K4907101), for the purpose of improving the folding stability so that both KpSP and GFP can function properly (3). Additionally, to enhance the efficiency of induction, E. coli DH10β; was selected as the expression strain for the following characterization.
Test
The test method was the same as which in Cycle 1. At this time, the bacteria expressing KpSP-GFP showed bright fluorescence upon excitation, and the normalized fluorescence intensity was much higher than that of the negative control (Fig. 4a), which meant the correct folding structure of GFP was retained. Moreover, the fluorescence intensity ratio of supernatant to the bacterial culture of KpSP-GFP was much higher than that of those expressing GFP only in cytoplasm (Fig. 4b), representing the good performance of KpSP.
Learn
The result indicated that KpSP can function well to direct target proteins out of the engineered bacteria as expected, which lays the foundation of our project for anti-soil frost. Although the fluorescence was recovered, the normalized fluorescence intensity of the bacterial culture of KpSP-GFP was still lower than that of the intracellularly expressed GFP, which meant the negative effects were not eliminated completely. From another aspect, the lesson implied an important but easy-to-be-neglected issue between the folding properties of target proteins and the (potential) transmembrane mechanisms of the implemented secretion systems (Fig. 4c), which might be helpful and valuable for other teams or labs when they are considering ideas about the secretion of recombinant proteins from engineered bacteria.
Conclusion
Engineering a biological system cannot be easy, we need to pay our all attention to it and move towards success through experiments and feedback step by step. In the beginning, we found that the KpSP signal peptide might influence the folding or maturation of GFP negatively when they are directly fused. For optimizing, we added a linker between KpSP and GFP, and then validated their function successfully. Fig. 5 demonstrates how we change the coding sequence of KpSP-GFP fusion protein to get the functional KpSP-GFP protein through successive iterations of the DBTL (Design-Build-Test-Learn) cycle.
Section 2: Harnessing pEcCas/pEcgRNA system to eliminate the cryptic plasmids of E. coli Nissle 1917
Introduction
For bacterial cellulose (BC) production, after comprehensive consideration, we chose E. coli Nissle 1917 (EcN) as our engineered bacteria, which will then be co-cultured with E. coli BL21(DE3) for the production of water-retention material, BC/HA (see Design page for more details). During the project, some modifications for the chassis EcN, especially the elimination of endogenous cryptic plasmids, have been achieved for the purpose of improving the production of BC.
Cycle 1: Construction of a circuit for characterizing the growth rate of E. coli Nissle 1917
Design
The water-retention material BC/HA that we plan to produce is a cross-linked product of BC and HA, and the ratio of BC to HA in the product will directly affect the water-holding capacity of the material. In order to ensure that the BC/HA we produce has the best water retention performance, we decided to regulate the strain ratio of two engineered bacteria, E. coli Nissle 1917 and E. coli BL21(DE3), during the process of co-culture fermentation. Inspired by previous work, we planned to introduce different fluorescent proteins into the two engineered bacteria respectively and use the ratio of the two fluorescence intensities to reflect the ratio of the two strains (4). Specifically, we planned to express the red fluorescent protein in E. coli Nissle 1917 and hence constructed a corresponding circuit (Fig. 6).
Build
There were four basic parts assembled into the vector pSB3C5: promoter (BBa_J23100), RBS (BBa_B0034), rfp coding sequence (BBa_K4907037), and terminator (BBa_B0015). We got the constructed plasmid successfully and transformed it into E. coli Nissle 1917.
Test
Surprisingly, after several trials, we found that our constructed plasmid, J23100-B0034-rfp-B0015_pSB3C5 could not be successfully transformed to E. coli Nissle 1917. At the same time, we also tried to transform another plasmid into E. coli Nissle 1917, including B0034-hasA-B0015_pSB1C3, J23100-B0034-gfp-B0015_pET-28a(+), pT7-B0034-rfp_pET-28a(+) and J04450_pSB4A5. We found those plasmids that are based on the pET-28a(+) or pSB4A5 vector are transformable, while the others are not transformable (Fig. 7).
Learn
Through the results, we hypothesized that there might be some endogenous plasmids in the E. coli Nissle 1917 strain, which have conflictive replication mechanisms with that of the pSB1C3 and pSB3C5 vectors that we are going to transform into. It is reported that two constitutive plasmids, pMUT1 and pMUT2, generally exist in E. coli Nissle 1917 and some called them cryptic plasmids (5). The presence of these two constitutive plasmids will affect the transformation of exogenous plasmids, i.e., plasmid incompatibility occurs (5). To better facilitate our subsequent engineering and experimental operations, we decided to explore whether these cryptic plasmids existed in the E. coli Nissle 1917 strain used in our lab or not.
Cycle 2: Confirmation that the E. coli Nissle 1917 strain we used contains pMUT1 with pMUT2
Design and Build
We got the sequences of two plasmids from NCBI (pMUT1: NZ_MW240712, pMUT2: NZ_CP023342) and designed the corresponding primers according to the previous work (6). Meanwhile, we searched for suitable restriction sites on the plasmid in order to linearize them and confirm the size of the plasmids (Fig. 8).
| Primer | Sequence (5′→3′) | Notes |
|---|---|---|
| pM1S2chk_F | GAATAGGGTGACACTGGCGCC | Forward primer for pMUT1 |
| pM1S2chk_R | CCAGATGGCATTGTAACAGACTTCTC | Reverse primer for pMUT1 |
| pM2S3chk_F | GATAAAACTATCAACTCACCGTCTTG | Forward primer for pMUT2 |
| pM2S3chk_R | GCCGTTGGTCTTTACTGATTTTAAG | Reverse primer for pMUT2 |
Test
Firstly, we extracted the plasmids from E. coli Nissle 1917 and performed PCR using the designed primers (Table 1), followed by agarose gel electrophoresis. At the same time, we performed restriction digest of the extracted plasmids (pMUT1 at the Nde I site and pMUT2 at the EcoR I site). The bands' sizes were consistent with the predicted ones generated from the known maps of pMUT1 and pMUT2, whatever the results of PCR (Fig. 9a) or digestion (Fig. 9b), which indicated that the E. coli Nissle 1917 we used did contain those two constitutive plasmids, pMUT1 and pMUT2.
Learn
From the experimental results, we confirmed that the E. coli Nissle 1917 strain in our lab contained the plasmids pMUT1 and pMUT2. In order to facilitate our subsequent experimental manipulations, we decided to eliminate these two plasmids.
Cycle 3: Elimination of pMUT1 using the pEcCas/pEcgRNA system
Design
The pEcCas/pEcgRNA system (Fig. 10a) was modified from pCas/pTargetF, a tool that has been shown to have higher genome editing efficiency in E. coli (7). In this system, pEcCas (Addgene number: 73227) harbors a sequence that expresses the Cas9 protein, while the sgRNA sequence expressed by pEcgRNA (Addgene number: 166581) will bind to the Cas9 protein to form a surveillance complex. Once the surveillance complex finds the target sequence, it cleaves the DNA double strand to cause DSB (double strand break) (8, 9). Considering the performance of pEcCas/pEcgRNA system and higher elimination efficiency of pMUT1 reported by previous research (6, 10), we therefore performed pMUT1's elimination at first. The sgRNA sequence, or specifically N20 sequence, will be designed to target a specific sequence on pMUT1 in this case (Fig. 10b).
Build
We first transformed the pEcCas plasmid into E. coli Nissle 1917. After successful transformation, the bacteria were cultivated and made into electrocompetent cells. Meanwhile, we designed the N20 sequence targeting pMUT1 on the online server (CRISPR) and constructed the pEcgRNA plasmid targeting pMUT1 according to the methods of the original paper(7). Finally, we transformed the constructed pEcgRNA-pMUT1 plasmid into E. coli Nissle 1917 harboring pEcCas by electroporation. (For more details, please see the protocol: Harnessing the pEcCas/pEcgRNA system for the elimination of cryptic plasmids in E. coli Nissle 1917)
Test
| Primer | Sequence (5′→3′) | Notes |
|---|---|---|
| pMUT1_F | ATGAGCTATGCGATAACTTTAACTGTGAAGC | Forward primer for pMUT1 |
| pMUT1_R | ATACTTTGCCAGTCCAAGCTGGGAT | Reverse primer for pMUT1 |
The transformants were randomly picked and verified via colony PCR. Next, we extracted the plasmids of those colonies that showed no target bands of colony PCR, and performed another PCR to verify whether the pMUT1 was really eliminated or not (primers are listed in Table 3). Luckily, there is one colony showing no bands after the second round of PCR, which indicated that the elimination of pMUT1 was achieved (Fig. 11, Lane 1).
Then the pEcCas/pEcgRNA plasmids of the chosen colony were cured, followed by the preparation of EcN ΔpMUT1 chemically competent cells. Subsequently, we tested whether I0500-B0034-tmafp-B0015_pSB1C3 (Fig. 12) and J23100-B0034-gfp-B0015_pSB3C5 can be transformed into EcN ΔpMUT1 or not, and finally found that the plasmid based on pSB3C5 backbone was still not transformable.
Learn
After the elimination of pMUT1, pSB1C3-bone plasmids could be transformed into E. coli Nissle 1917, which indicated that pSB1C3 might have a similar replication mechanism with pMUT1, and this was also consistent with the conclusion of the previous work (10). Due to the failure of transforming the pSB3C5-bone plasmids, we hypothesized that the presence of pMUT2 would still affect the transformation of the exogenous plasmid, and therefore we continued our work on the elimination of pMUT2 from EcN ΔpMUT1.
Cycle 4: Elimination of pMUT2 using the pEcCas/pEcgRNA system
Stage 1
Design and Build
Similarly, we designed the N20 sequence and constructed the pEcgRNA-pMUT2 plasmid (Table 4), and then transformed it into EcN ΔpMUT1 harboring pEcCas (Fig. 10c).
| Sequence (5′→3′) | Notes |
|---|---|
| ATTACCTGATCCTTCCGCAG | N20 sequence |
| TAGTATTACCTGATCCTTCCGCAG | Top strand sgRNA (pMUT2) with overhang sequence |
| AAACCTGCGGAAGGATCAGGTAAT | Bottom strand sgRNA (pMUT2) with overhang sequence |
Test
| Primer | Sequence (5′→3′) | Notes |
|---|---|---|
| 8F | AAAGAGTCCTTAGCACCCGC | Forward primer for pMUT2 |
| 8R | GGTGCGGTGTCAAGAAAGTT | Reverse primer for pMUT2 |
Learn
Consistent with most of what has been noted in previous research, the elimination of pMUT2 often requires a lot of effort (6, 10). Since the pMUT2 plasmid could not be eliminated after many times of trials and failures, we decided to give up our current scheme and search for other feasible approaches. It was reported that the relE-relB toxin-antitoxin system is naturally present on pMUT2, and the introduction of the expression cassette of relB into the plasmid harboring the sgRNA designed to target the relaxase coding sequence on pMUT2 would contribute to eliminating pMUT2 (6). We therefore decided to follow this experience for pMUT2's elimination.
Stage 2
Design and Build
We obtained the sequence of relB and the N20 sequence targeting the relaxase coding sequence on pMUT2 from the article (6). Next, we amplified the fragment comprising the coding sequence of relB from pMUT2 and performed overlap-extension PCR (OE-PCR) by using the amplified fragments and the synthesized upstream untranslated sequences as templates to obtain the expression cassette of relB. Subsequently, we cloned the expression cassette of relB into pEcgRNA-relaxase that was constructed before (Table 6) via Gibson assembly to obtain the construction of the final sgRNA-harboring plasmid. (For more details, please see the protocol: Harnessing the pEcCas/pEcgRNA system for the elimination of cryptic plasmids in E. coli Nissle 1917)
Test
In previous trials of validations, we picked the colonies and then performed colony PCR, followed by plasmid extraction and another round of PCR. However, at this time, we found that the primers we designed were not specific enough, and hence they can bind to pEcCas and pEcgRNA, leading to false-positive results. Therefore, we decided to implement agarose gel electrophoresis of the extracted plasmids directly instead of colony PCR, after curing the pEcgRNA-relaxase plasmid harboring the expression cassette of relB. The extracted plasmids from the cured transformants showed no bands of pMUT2 (Fig. 14), which validated the elimination of pMUT2.
After that, we tried to transform J23100-B0034-gfp-B0015_pSB3C5 into the E. coli Nissle 1917 strain that contained no endogenous plasmids (EcNP for short), and the results showed that the pSB3C5-bone plasmid could be successfully transformed eventually (Fig. 15).
Learn
We successfully eliminated pMUT2 and the experimental results showed that there is no obstacles for transforming the plasmids we want, which would facilitate our downstream experimental manipulations. In addition, since we have already obtained wild type EcN, EcN ΔpMUT1 that eliminated pMUT1, as well as the EcNP that eliminated both pMUT1 and pMUT2, we were interested in the growth rate and the ability to express proteins of the three strains, and hence we carried out the experiments of next stage.
Cycle 5: Testing the performance of EcN, EcN ΔpMUT1, EcNP
Design and Build
We planned to express red fluorescent protein in three strains, EcN, EcN ΔpMUT1 and EcNP, so as to compare the growth rate and protein expression level of these three strains by comparing the OD600 and the fluorescence intensity. Thus, we reconstructed a gene circuit J23100-B0034-rfp on pET-28a(+).
There were four basic parts assembled into the vector, pET-28a(+): promoter (BBa_J23100), RBS (BBa_B0034), and rfp coding sequence (BBa_K4907037). We got the constructed plasmid (BBa_K4907037)_pET-28a(+) successfully and transformed it into EcN, EcN ΔpMUT1 and EcNP, respectively.
Test
We conducted OD600 and fluorescence intensity measurements on the three strains which harbored J23100-B0034-rfp_pET-28a(+).
Learn
The growth rate of EcNP containing no endogenous plasmids was faster than the other two strains (Fig. 16a), while the expression level of fluorescent protein showed no significant differences among the three strains (Fig. 16b), which indicated the elimination of pMUT1 and pMUT2 might have little contribution to improve the expression of heterogenous proteins in E. coli Nissle 1917.
Conclusion
Firstly, we found plasmid incompatibility during our experiments, then, leveraging the pEcCas/pEcgRNA
system with several rounds of attempts and modifications of protocols, we finally eliminated the two
cryptic plasmids, pMUT1 and pMUT2, from the E. coli Nissle 1917 strain we used. We found
that EcNP had a higher growth rate but a similar protein expression level compared to the wild
type EcN. The elimination of the cryptic plasmid increases the space for the exogenous plasmids,
which undoubtedly provides a better performing engineered bacteria for our subsequent experimental
operations as well as our project. Additionally, it was the first time that the pEcCas/pEcgRNA
system which is commonly used for genome editing was implemented and demonstrated to eliminate
endogenous plasmids, which would extend the boundary of applications of this pEcCas/pEcgRNA
system. Here, we have organized the protocol for the whole plasmid elimination workflow,
with an expectation to offer valid help to other teams or labs which are using or want to use
E. coli Nissle 1917 for various synthetic biology applications and contribute
to the iGEM community.
All in all, engineering a biological system cannot be easy, we need to put our heart into it and move towards success through experiments and feedbacks step by step. Here, we have tried to demonstrate how we harnessed the pEcCas/pEcgRNA system to eliminate the cryptic plasmids of E. coli Nissle 1917 through successive iterations of the DBTL (Design-Build-Test-Learn) cycle (Fig. 17).
Reference
- Y. Cui et al., Efficient secretory expression of recombinant proteins in Escherichia coli with a novel actinomycete signal peptide Protein Expression Purif. 129, 69-74 (2017).
- T. D. Craggs, Green fluorescent protein: structure, folding and chromophore maturation. Chem. Soc. Rev. 38, 2865-2875 (2009).
- C. J. Crasto, J.-a. Feng, LINKER: a program to generate linker sequences for fusion proteins. Protein Eng. Des. Sel. 13, 309-312 (2000).
- J. Gutiérrez Mena, S. Kumar, M. Khammash, Dynamic cybergenetic control of bacterial co-culture composition via optogenetic feedback. Nat. Commun. 13, 4808 (2022).
- Y.-J. Lan et al., Development of Escherichia coli Nissle 1917 derivative by CRISPR/Cas9 and application for gamma-aminobutyric acid (GABA) production in antibiotic-free system. Biochem. Eng. J. 168, 107952 (2021).
- A. Kan, I. Gelfat, S. Emani, P. Praveschotinunt, N. S. Joshi, Plasmid vectors for in vivo selection-free use with the probiotic E. coli Nissle 1917. ACS Synth. Biol. 10, 94-106 (2020).
- Q. Li et al., A modified pCas/pTargetF system for CRISPR-Cas9-assisted genome editing in Escherichia coli. Acta Biochim. Biophys. Sin. 53, 620-627 (2021).
- K. Wang, Y. Wang, Basic properties and research methods of a new CRISPR-Cas system: CasX as an example. Microbiology China 48, 1227-1238 (2021).
- F. Jiang, J. A. Doudna, CRISPR-Cas9 structures and mechanisms. Annu. Rev. Biophys. 46, 505-529 (2017).
- F. Ba et al., Expanding the toolbox of probiotic Escherichia coli Nissle 1917 for synthetic biology. Biotechnol. J. n/a, 2300327 (2023).
