"Strainer" is an efficient and versatile tool, and we have demonstrated in the lab that it can be used well in the following aspects.

The Gene Inactivation Experiments

First of all, we designed four gRNAs (in the EC21 to EC24 plasmids) targeting on the different positions of the galK gene and the related donor DNAs contained the stop codon (TAA). In addition, each gRNA with its related donor DNA was constructed in the same editing plasmid (Fig. 1A), which is good for genome editing using gRNA libraries in future. If the stop codon is inserted into the galK gene, white colonies will be observed on the MacConkey agar (Fig. 1B). Otherwise, the red colonies (the unedited cells) will be observed on the MacConkey agar (Fig. 1B).

We used two E. coli strains (EC85 and EC88) with the plasmid harboring lambda red system and different CRISPR proteins (MAD7 and AsCas12a) as the host strains to test our four editing plasmids. We found that the editing efficiency using EC21 plasmid is the lowest in both host strains (Fig. 1C, D).

Fig. 1. The gene inactivation experiments. (A) Transformation of EC21, EC22, EC23, and E24 plasmids into E. coli which contains Cas and λred genes, respectively. (B) Two MacConkey agar plates. One is full of the red colonies (the unedited cells), the other is full of the white colonies (the edited cells). (C) The CFU and editing efficiency after transformation of the four gRNA plasmids into EC85 strain, which is MG1655 with pM7. (D) The CFU and editing efficiency after transformation of the four gRNA plasmids into EC88 strain, which is MG1655 with pAs.

We designed another 4 gRNAs, which contained the promoter sequences of the different SOS genes (lexA, sula, umuD, and recA genes). The spacer sequences of these gRNAs are the same, which is targeting on the resistance genes (kanamycin) of the plasmid harboring sacB gene. After that, we used the EC21 plasmid as a template to build 4 new plasmids (LexA, Sula, UmuD, and RecA) using gibson assembly method (Fig. 2A). We firstly tested the new editing plasmid LexA using lexA gene promoter in EC85 to validate the “strainer” method. At the 0.1% sucrose concentration, the highest CFU/μg and the highest editing efficiency were found in the experiments. The editing efficiency using the LexA plasmid is 25% higher than the control without using the “strainer” method. However, the CFU/μg decreased significantly compared to the control group that did not use “strainer” (Fig. 2B).

Fig. 2. Preliminary verification of the “strainer” method. (A) Transformation of the recombinant plasmids into EC85 separately. (B) The editing efficiency and CFU/ug in the different sucrose concentrations.

We then optimized the sucrose concentrations to further improve CFU/μg while keeping the high editing efficiency. By using the different concentrations of sucrose, we found that the CFU/μg increase 6-fold using 0.01% sucrose, while the editing efficiency was consistent with that of 0.1% sucrose (Fig. 3A). However, the CFU/μg was still 96% lower than that of the control (Fig. 3A). To this end, we tried to further improve CFU/μg by designing and modifying SacB protein to reduce its toxicity. In addition, we used 0.01% sucrose to tested another three new editing plasmids. All of them showed that the editing efficiency was ~25% higher than the control without using the “strainer” method (Fig. 3B).

Fig. 3. Optimization of sucrose concentrations. (A) Sucrose concentration gradient experiments. (B)The editing efficiency and CFU/ug of all the four new editing plasmids under 0.01% sucrose condition.

Improve The CFU/μg of “strainer” Method with SacB Mutant

We designed the SacB_S164T mutant which showed less toxicity for E. coli using Molecular dynamics (MD) simulations. Thus, we used Mut Express II Fast Mutagenesis Kit to construct the plasmid with the SacB_S164T mutation. Then, we tested the SacB_S164T mutant for the “strainer” method in EC85. The results showed that the editing efficiency of SacB_S164T mutant was 25% higher than the control without using “strainer” method. The CFU/μg of SacB_S164T mutant increased ~3-fold compared to the original “strainer” method while still keep the high editing efficiency (Fig. 4A).

We also used the same condition to test SacB_S164T mutant in EC88. The results showed that the editing efficiency using “strainer” method with SacB_S164T mutant was 4-fold higher than the control, although the CFU/μg of SacB_S164T mutant was still 71% lower than the control without using the “strainer” method (Fig. 4B). These results showed that our “strainer” method worked as a good purification system to remove the unedited cells.

Fig. 4. Improve the CFU/μg of “strainer” method with SacB mutant. (A) The CFU/μg and editing efficiency of the control group, SacB and SacB_S164T under 0.01% sucrose condition. (B) Used 0.1% and 0.2% sucrose concentration to test SacB _S164T in EC88.

Long Gene Fragment Insertion

SS9 is a safe locus in the E. coli genome, where we can insert the long gene fragments without a strong effect on the growth of E. coli. Therefore, we chose this region to verify the "strainer" method for the long DNA fragment insertion experiments.

We used the “strainer” and the original CRISPR/Cas method to insert the isopropanol (IPA) production pathway genes into the E. coli genome. Through the colony PCR method, we found that the insertion efficiency using the original CRISPR/Cas method was 3%, while the insertion efficiency using the “strainer” method reached 33% (Fig. 5A). Obviously, our “strainer” method had unparalleled advantages for the long DNA fragments insertion.

Fig. 5. Long gene fragment insertion and “strainer” method utilization to improve the editing efficiency for IPA-producing strain construction. (A) The editing efficiency of long gene fragment insertion by using CRISPR/Cas9 and “strainer” methods. (B) The editing efficiency of introducing the FliA_R94K mutation by using CRISPR/Cas9 and “strainer” methods. (C) IPA production of EC61 and EC65 strains.

Use “strainer” Method to Improve The Editing Efficiency For IPA Producing Strain Construction

We named the E. coli strain with IPA production pathway genes as EC61. Then, we then further modified EC61 strain by introducing the FliA_R94K mutation conferring the IPA resistance. We used the “strainer” and the original CRISPR/Cas methods to introduce the mutation into EC61. The “strainer” method showed 20% higher editing efficiency than that of the original CRISPR/Cas method (Fig. 5B).

After that, we used EC61 and EC65 to carry out IPA fermentation experiments. From the results, we could see that the IPA titer of EC65 was 3-fold higher than that of EC61 (Fig. 5C).

In the whole process of the strain modification for IPA production, our “strainer” method could significantly accelerate the construction speed of the strains.