Expression of CRISPR/Cas9 with gRNA targeting the T7 phage genome increased the phage defense ability of E. coli BL21(DE3)
To determine whether programed CRISPR/Cas9 can increase the ability of BL21 to defend itself against bacteriophages, we designed three different N20s, respectively targeting the genes of the tail tubular protein gp12, capsid assembly protein and 3.8 protein in the T7 phage genome (Fig. 1). Four strains, BL21(pcas9, pTgRNA), BL21(pcas9, pCgRNA), BL21(pcas9, p3.8gRNA), and BL21(pcas9, p3gRNA) were constructed, among which plasmid p3gRNA contained gRNA of all three loci. We prepared phage stock solution and 1μL of it contains about 1010 phage. At MOI 0.02, the results showed that the OD600 of BL21(pcas9, pTgRNA) and BL21(pcas9, p3gRNA) was significantly higher than that of BL21(pCas9) with phage infection (Fig. 2). However, the OD600 of BL21(pcas9, pCgRNA) and BL21(pcas9, p3.8gRNA) was not distinctly higher than that of the control (Fig. 2). These results indicated that programed CRISPR/Cas9 targeting the phage genome increased the ability of BL21 to defend itself against phage infection, and the targeted loci played a critical role.
Optimization of the programmable CRISPR/Cas9 phage defense system
To better control the expression of CRISPR/Cas9, we replaced the constitutive promoter with the arabinose-inducible promoter (pBad promoter), resulting in strains BL21(pTFG025, pTgRNA), BL21(pTFG025, pCgRNA), BL21(pTFG025, p3.8gRNA) and BL21(pTFG025, p3gRNA). The OD600 of all the engineered strains after phage infection was higher than that of the control, and slightly higher than that of the strains with the constitutive promoter (Fig. 3A). The results illustrated that gene expression using the pBad promoter improved the performance of the CRISPR/Cas9 defense system, and suggested that inducible expression could be a better choice.
Considering that the BL21 series strains contain T7 polymerase, we replaced the pBad promoter with the T7 promoter to improve the expression of the CRISPR/Cas9 defense system. By changing the promoter of the Cas9 plasmid, strains BL21(pT7-cas9, pTgRNA), BL21(pT7-cas9, pCgRNA), BL21(pT7-cas9, p3.8gRNA) and BL21(pT7-cas9, p3gRNA) were constructed. However, when the concentration of the inducer IPTG was at 0.4 mM, the performance of the CRISPR/Cas9 defense system was not as good as that of the system expressed using pBad (Fig. 3B). To better modulate the expression of the CRISPR/Cas9 genes, the constitutive promoter for gRNA transcription was also replaced by the T7 promoter, and the IPTG concentration was reduced to 0.05 mM. The newly constructed strains had a very strong defense against phage infection, that the OD600 of BL21(pT7-cas9, pT7-TgRNA, prfp) and BL21(pT7cas9, pT7-3gRNA, prfp) after 4 hours of phage infection was significantly improved to reach 2.0, which was similar to the control culture with no phage infection and obviously higher than the OD600 of BL21(pTFG025, p3gRNA) (Fig. 3C). To determine its performance with more severe phage infection, the phage concentration was increased 10-fold. Under such conditions, BL21(pT7cas9, pT7-3gRNA, prfp) still maintained an OD600 of around 1.7, slightly lower than that of the control (Fig. 3D) and the BL21(pT7cas9, pT7-3gRNA, prfp) in MOI 0.02, suggesting that the CRISPR/Cas9 defense system was robust in the presence of different concentrations of phage particles. After 4 hours the defensive ability of CRISPR/Cas9 systems were not as obvious as that at early time points. At 16h time points, the difference among different CRISPR/Cas9 systems were also reduced. However, the growth condition of our best engineered strain was still significantly better than the infected control strain.
We also integrated the pT7cas9 and pT7-3gRNA into the chromosome of E. coli BL21 at poxb locus by pCAGO[26] to construct the strain BL21-T7-3gRNA-T7-Cas9. However, the resistance of the integration strain against phage infection was not as good. The growth curve of BL21-T7-3gRNA-T7-Cas9 (prfp) was similar to the control with no CRISPR/Cas9, and significantly lower than that of the plasmid-based strain BL21(pT7cas9, pT7-3gRNA, prfp) (Fig. 3E). SDS-PAGE was used to analyze the expression of the Cas9 protein, which revealed that its content in BL21-3gRNA-cas9 was very low (Fig. S1). The low expression may be due to only one chromosomal copy of the CRISPR/Cas9 gene being present in the integrated strain.
Furthermore, we performed a colony count to quantify the actual living cells in the samples, in addition to the OD600 readings. When BL21 was infected with the T7 phage, there were almost no surviving cells after either 4h or 16h of culture. By contrast, the control sample with no infection had a normal density of around 108 to 109 cells/mL. The strain BL21(pT7cas9, pT7-3gRNA, prfp) had a significantly higher survival rate, with 4.8×107 and 2.7×107 viable cells per mL after 4 hours of infection with 1μL and 10μL of phage stock, respectively (Table 2, Fig.S2). The viable cell numbers of both samples were only one magnitude lower than that of the strain with no infection, further proving the efficacy of the CRISPR/Cas9 phage defense system.
Increased protein expression of BL21 with the CRISPR/Cas9-based phage defense system
In order to evaluate the protein production capacity of BL21 stains with the CRISPR/Cas9 phage defense system, the plasmid prfp was transferred into the strains with CRISPR/Cas9 induced by T7 promoter. This plasmid was derived from the broad-host-range plasmid pBBR1MCS2, and the red fluorescent protein (rfp) gene driven by the constitutive promoter BBa_J23100[27]. Rfp was expressed in the strains and the fluorescence value was measured after phage infection (Figs. 4A and 4B). For each strain, the fluorescence value of the vector control (pΔrfp) is about 0. It was found that all engineered strains performed better than the control strain with no CRISPR/Cas9. However, they were not as good as the control with no phage infection. The best performing strain was still BL21(pT7cas9, pT7-3gRNA, prfp) as in the OD600 experiments, which had a fluorescence value about 60% that of the un-infected control (Fig.S3). We also calculated the value of FV/OD600. However, the value of all of the strains with the system was lower than the control strain, which indicated that the rfp gene might also be persistently translated when control strain was dead. The results support our hypothesis that the CRISPR/Cas9 defense system could be employed to protect BL21 and increase its protein production in bioprocesses with intractable phage infections in the equipment.