Design

Overview

We design CRISPRa and CRISPRi systems to evaluate and select dCas9 mutants generated by random mutagenesis in order to obtain THE ONE that has both high on-target level and low off-target level. We also make rational design with the assistance of model.

CRISPRa (CRISPR activation)

We read series of related papers, finding that there is an interesting Cas9-mediated regulating system, which can be remodeled to a new circuit to detect on-target rate of target sites^[1].

As the foundation of our work, we first construct a CRISPRa system composed of two parts, one of which contains a dCas9 protein with an RNA polymerase (RNAP) ω subunit fused to its C-terminal, while the other includes an RFP reporter gene with a weak promoter J23117 as well as a protospacer 96nt upstream of the transcription start site, and an sgRNA construction module. When dCas9 binds with protospacer, the ω subunit on dCas9 could recruit RNA polymerase holoenzyme, facilitating the transcription of downstream reporter gene. We set the distance between the protospacer and promoter of reporter gene refering to preliminary experiments^[1].

Distance between the protospacer and promoter of reporter gene.

For the convenience of our serial fragments insertion, we designed two plasmids, pdCas9 and pScreen. The first plasmid part pdCas9 is generated by adding the dCas9 with a constitutive promoter to the plasmid vector pACYC184, which has chloramphenicol resistance marker and low copy number. The second plasmid part pScreen contains the sgRNA construction module, the target-RFP sequence and the ampicillin resistance marker from high-copy-number vector pUC57.

With these two plasmids, we can conduct a simple on-target detection. Since there are considerable copies of the target-RFP part, the red fluoresce intensity can effectively report the on-target level of a single cell, which is reasonable for flow cytometry selection.

Furthermore, after we find out the latent off-target sequences by literature searching and modeling, we can modify the on-target module to an off-target module by changing the protospacer sequence. We also replace the reporter in off-target module by an eGFP gene before piling these modules up in a pUC57-sgRNA plasmid for a dual detection system.

CRISPRi (CRISPR interference)

Under the guidance of Professor Xiao Yi, we learned that the system of CRISPR interference (CRISPRi) also had a good regulatory effect. Based on our original transcriptional activation gene circuit, we designed a transcriptional inhibitory gene circuit. The dCas9 provides a simple and robust technology for gene repression and activation, and can target almost any DNA sequence aided by sgRNA. CRISPRi inhibits transcription by sterically blocking the RNA polymerase ^[2].

Mechanism of CRISPR-mediated transcriptional inhibition.

Our inhibition system also contains two plasmids, pdCas9 and pIn. The first plasmid part pdCas9 is the same as what we use in our CRISPRa system. The second plasmid part pIn contains the sgRNA construction module, the target-RFP sequence and the ampicillin resistance marker from high-copy-number vector pUC57.

Circuit design for CRISPRi screening plasmid.

In our system, the expression of RFP reporter gene is under control of pBAD, which could be induced by arabinose. And the target sequence is located directly downstream of the promoter, where dCas9 could bind and block RNAP. Because of the presence of the constitutive promoter in plasmid pdCas9, dCas9 is continuously expressed throughout the growth stage of bacteria and has a higher likelihood of identifying and binding to targets. Therefore, a relatively weak fluorescence expression of bacteria indicates a dCas9 with higher on-target rate that effectively inhibits the expression of reporter gene.

Furthermore, we can modify the on-target module to an off-target module by adopting the same pattern, in which low green florescence characterization indicates off-target. By using these parts, we construct a plasmid with both target and off-target sites, which reflect the on-target and off-target of dCas9 by fluorescence simultaneously. Thus we can use this system to screen for evolved dCas9.

Circuit design of screening plasmid with sgRNA construction module.

Meanwhile, we have made another lure design that we just insert the lure sequence to the pIn plasmid. Since the lure can compete with target for dCas9 binding, the decline extent in RFP intensity reflects the target rate of lure, which is also called the off-target rate. Though there is only one output of this simplified system, it's enough for detecting dCas9's overall performance.

Another screening plasmid design with sgRNA construction module

Mutagenesis library construction

Mutation library is crucial for directed evolution and its construction can be realized by various methods, such as error-prone PCR, DNA shuffling, mutator strain, etc. After comparing these methods carefully, we choose error-prone PCR and DNA shuffling as our construction methods. We carry out error-prone PCR at first, conduct selection through several rounds of screening, and then apply DNA shuffling to distribute mutated sites randomly.

Mechanism of Error-Prone PCR and DNA shuffling

Screening

After obtaining the mutagenesis library, we co-transform mutant dCas9 and pScreen (or pIn in CRISPRi) into MG1655 to construct the dual detection system. Given that there are multiple version of mutated dCas9-ω, considerable cell amount and reporter to show on-target/off-target level in our sample, we can sort cells with both high on-target level and low off-target level, indicated by bright red fluoresce and dim green fluoresce in CRISPRa (or dim red fluoresce and bright green fluoresce in CRISPRi). Though biological random variation exists in the system, we still get a better dCas9 protein through each flow cytometry sorting.

We will also use DNA shuffling technology, which can combine mutageneses together, providing further efficiency of our detection system. We will conduct 2 to 3 rounds of screening after each error prone PCR to screen out the most possible forward mutants. Then DNA shuffling is carried out to accumulate the advantages of mutants, so as to maximize the utilization of the mutant library.

After many cycles of mutation, selection and hybridization, advantageous mutations are accumulated constantly in our library, eventually achieving the effective directed evolution and obtaining our dreamed protein.

Rational Design

We also apply rational design to optimize our evolution progress. Our software, PROMARE finds the key residues about dCas9 on-target rate, and site-directed mutageneses are applied to test the model and accelerate the evolution. Promare is a protein design software developed by our team. It can help users find the key residues in protein and design a strategy to improve protein function.

First, we extracted and compared the distances between DNA-RNA recognition bases and its neighbor amino acids. Compared the distances from spCas9 and xCas9, we find the difference between spCas9 and xCas9 in top 10 closest protein-nucleotide distances. Combined with the molecular dynamics results, we know that the distances play a crucial role in off-target effects.

So we carried out a saturated mutation on all amino acid have contact with nucleotides. The same distance features were calculated and compared. Therefore we selected those mutation site with highest potential to affect the off-target effect to do saturation mutation experiment.

References

[1] Bikard, D., Jiang, W., Samai, P., Hochschild, A., Zhang, F., & Marraffini, L. A. (2013). Programmable repression and activation of bacterial gene expression using an engineered CRISPR-Cas system. Nucleic acids research, 41(15), 7429–7437. https://doi.org/10.1093/nar/gkt520
[2] Yao, R., Liu, D., Jia, X., Zheng, Y., Liu, W., & Xiao, Y. (2018). CRISPR-Cas9/Cas12a biotechnology and application in bacteria. Synthetic and systems biotechnology, 3(3), 135–149. https://doi.org/10.1016/j.synbio.2018.09.004