home

overview

Result

In order to select dCas9 protein with higher on-target and lower off-target rate, we designed a screening system and two different judging methods by monitoring alteration of two kinds of fluorescence level.

Transcription activation

In this circuit, when dCas9 binds with target or lure sequence, the $\omega$ subunit will recruit RNA polymerase, initiating transcription of downstream reporter gene – RFP and eGFP, respectively. The design of transcription activation system is inspired by a published paper, with target or lure sequence located 59bp upstream of the -35 element. Target and potential off-target sequence (lure) of PDCD1 are taken from previous paper as well as predicted by our own software.

Construct plasmid for verifying transcription activation

To verify the feasibility of our system, we first constructed a module with only target sequence and each reporter gene. J23117 promotor and 35bp spacer sequence were obtained through PCR, which were then ligated to target sequence by designing target sequence on the primer. We used overlap PCR to ligate the target+spacer+J23117 with RFP and eGFP separately.

Distance between the protospacer and promoter of reporter gene.

RFP fragment with terminator was obtained from pE1a-RFP kindly offered by Tongji-China. Since our team did not have available eGFP gene with terminator, we conducted the procedure below to add a double terminator to eGFP. Due to complex secondary structure of double terminator, we gave up the idea to perform PCR twice to add terminator to eGFP fragment. Rather, we first obtained terminator by performing PCR on pE1a-RFP and then link target+spacer+eGFP and terminator through overlap PCR. Success was proven by both gel electrophoresis of PCR product and DNA sequencing.

PCR product of target+spacer+eGFP+terminator . (Lane 1: 2000 bp marker;Lane 2-9: overlap PCR product)

Circuit of target+spacer+J23117+RFP/eGFP+terminator.

Gibson homologous regions were added to this fragment through another PCR process.

We then linearized the backbone of pUC57 for Gibson assembly.

Sequentially, we assembled pUC57 backbone and the PCR product with Gibson assembly. The assembled product was transformed into DH5α. Colony PCR was conducted for positive colony selection and verification. And we named this plasmid for verifying transcription activation effect as pScreen-RTr/g(r for RFP while g for eGFP).

Verification of pScreen-RTg assembly through colony PCR. (Lane 1-8 and Lane 10-17: result of PCR product; Lane 9: 2000bpmarker. Lane 16 shows positive result.)

Verification of pScreen-RTr assembly through colony PCR. (Lane 2-25: result of PCR product; Lane 1: 2000bp marker. Lane 10 shows positive result.)

Construct plasmid for mutant libraries screening

After successfully constructed plasmid with only target and eGFP/RFP, we continued to add lures to this screening module. We first chose four lure sequences with highest off-target rate, two from literature and two from our software prediction (named lure 1/2/3/4).

We first tried Gibson assembly. Lure+eGFP fragment was obtained by conducting PCR and replacing target sequence with lure on the primer. Homologous fragments were added to lure+eGFP fragments.

Circuit of curcuit of Lure1/2/3/4+spacer+J23117+eGFP+terminator

Verification of pScreen-RTL1/3/4 assembly through colony PCR. (Lane 1-8: result of PCR product of pScreen-RTL1,Lane 2、3、7 show positive result; Lane 9-16: result of PCR product of pScreen-RTL3，Lane 11、14、15 show positive result; Lane 17-24: result of PCR product of pScreen-RTL4，Lane 17 shows positive result; Lane 25: 2000bp marker. )

Preparing $\Delta$rpoZ MG1655 strain

Since our team want to recruit RNA polymerase and initiate transcription of reporter gene through $\omega$ subunit-fusion dCas9 binding with target or lure sequence, innate $\omega$ subunit of E.coli must be removed to avoid disturbance. So we decided to knock out rpoZ - $\omega$ subunit encoding gene in E.coli, by homologous recombination of Lambda Red recombineering system.

pKD46 plasmid was transformed into E.coli strain MG1655 by preparing MG1655 electro-competent cells. Kanamycin resistant gene with homologous arm of rpoZ on MG1655 genome was obtained through PCR using pET28a plasmid as template.

L-arabinose was added to induce pKD46 gene expression of Exo, Beta, Gam protein, and Kana-homo fragment was transformed in to knock out rpoZ gene.

Knockout of rpoZ was selected and verified by using kanamycin selective medium and verifying the length of PCR product. We named this strain $\Delta$MG1655.

PCR product of Kanamycin resistant gene with homologous arm of *rpoZ* on MG1655 genome.(Lane 1, 3：result of PCR product. Lane 2：2000 Marker）

Verification of *rpoZ* deletion through colony PCR. (Lane 1: 5000 bp marker; Lane 2-25: result of PCR product. The length of PCR product of *rpoZ* is 464bp, while cells successfully knocking out *rpoZ* will have PCR product of 1155bp- the length of kanamycin resistant gene. Lane 3, 4, 10 and 21: positive colonies.)

System verification

After we finished the construction of pScreen-RTg and pScreen-RTr, we immediately verified the feasibility of our activation system. We prepared 5 kinds of strains carrying different plasmids and test their fluorescence intensity using flow cytometer for general screening and flowcytometry for single-cell screening.


Strain	pdCas9*	pScreen-RTg	pScreen-RTr
$\Delta$MG1655	-	+	-
$\Delta$MG1655	-	-	+
$\Delta$MG1655	+	+	-
$\Delta$MG1655	+	-	+
$\Delta$MG1655	-	-	-

*pdCas9 is generated by adding the dCas9-ω with a constitutive promoter to the plasmid vector pACYC184.

The results are shown below. Unfortunately, it seemed that dCas9-ω didn’t bring about transcription-activation effect for certain reason. We hypothesized that the force of ω-subunit recruiting RNP is too weak to activate transcription. Meanwhile, dCas9-ω binding to target may instead interfere with transcription as it acts as a “roadblock”.

Raw data: CRISPRa data

Transcription inhibition

In this circuit, target and lure sequence are designed right after promoters of reporter genes, RFP and eGFP, respectively. Binding and staying still of dCas9 prevents RNA polymerase from sliding by, thus inhibit transcription of downstream gene.

System verification

To validate the inhibition system can work, plasmids pE1a-RFP-sgRNA and pS8K-dCas9 kindly offered by Prof. Yi Xiao were transformed into MG1655 and underwent microplate photometer to see fluorescence difference.

MG1655 co-transformed with pE1a-RFP-sgRNA and pS8K-dCas9 were divided into four groups with three repeats in each group. The culture condition of each group is mentioned below. IPTG was added in the liquid culture to induce the expression of RFP, and L-arabinose was added to induce the expression of dCas9. Glucose was used to inhibit leaky expression of dCas9.


	1M arabinose stock solution	1M IPTG stock solution	0.5g/ml glucose stock solution
1	Added to liquid culture in a ratio of 1:200 at the start	Added to liquid culture in a ratio of 1:1000 when OD600 reaches 0.6	No
2	Added to liquid culture in a ratio of 1:200 when OD600 reaches 0.4	Added to liquid culture in a ratio of 1:1000 when OD600 reaches 0.6	No
3	No	Added to liquid culture in a ratio of 1:1000 when OD600 reaches 0.6	Added to liquid culture in a ratio of 1:250 when OD600 reaches 0.4
4	No	Added to liquid culture in a ratio of 1:1000 when OD600 reaches 0.6	No

The results are shown below, from which we could see 9.44 times of fluorescence/OD600 difference between control group (without any operation, only express RFP) and dCas9 expression group, indicating dCas9 do inhibit the expression of reporter gene.

Raw data: Transcription inhibition system verification

Construct plasmid for verifying transcription inhibition

Seeing the feasibility of the inhibition system, we started to construct our own plasmids.

Inducible araC and pBAD promoter was amplified from pS8k-dCas9 given by Prof. Yi Xiao. Our target sequence and restriction cutting sites were added to the fragment during PCR. In the meantime, target+RBS+RFP+term with restriction enzyme cutting site was amplified from pScreen-RTr.

The two fragments were ligated through overlap PCR, and the product was ligated into pUC57 using SacⅠ and EcoRⅠ. We named this plasmid for verifying transcription inhibition as pIn-RTr.

PCR product of araC+pBAD+target+RBS+RFP+terminator. (Lane 1: 5000 bp marker; Lane 2-13: overlap PCR product)

Circuit of araC+pBAD+target+RBS+RFP+terminator.

Verification of pIn-RTr assembly through colony PCR. (Lane 1-8、11-18: result of PCR product; Lane 9: 2000bp marker；Lane 10:5000bp marker. Lane 13-15 show positive result.)

Verifying the inhibition effect of dCas9 on pIn-RTr

We set up six culturing system (as shown in the table below) and used microplate spectrophotometer to examine fluorescence intensity. L-arabinose was added to induce the expression of RFP.


	pln-RTr	pdCas9	L-arabinose
MG1655	+	+	+
MG1655	+	+	-
MG1655	+	-	+
MG1655	+	-	-
MG1655	-	-	+
MG1655	-	-	-

The results are shown below, from which we could see 64.39 times of fluorescence/OD600 difference between transform group (without the expression of dCas9) and co-transform group(with the expression of dCas9), indicating our transcription inhibition system did work.

Raw data: Transcription inhibition system verification(pIn-RTr)

Construct plasmid for mutant libraries screening

Then we continued the construction of screening plasmid containing both target and lure with their reporter gene.

Inducible pBAD promoter was amplified from pS8k-dCas9 given by Prof. Yi Xiao. Our lure1/2/3 sequence and restriction cutting sites were added to the fragment during PCR. In the meantime, lure1/2/3+RBS+eGFP+term with restriction enzyme cutting site was amplified from pScreen-RTr.

The two fragments were ligated through overlap PCR, and the product was ligated into pUC57 using Pst1. We named this plasmid for screening transcription inhibition as pIn-RTL1/2/3. However, only lure3's overlap product is successfully connected. Due to the limited time, we decide to carry out screening experiment with lure3 first, and the rest will be completed later.

PCR product of pBAD+lure1+RBS+eGFP+term. (Lane 1: 2000 bp marker; Lane 2-10: overlap PCR product)

Verifying the inhibition effect of dCas9 on pIn-RTL3

We set up six culturing system (as shown in the table below) and used microplate spectrophotometer to examine fluorescence intensity.L-arabinose was added to induce the expression of RFP and eGFP.


	pln-RTL3	pdCas9	L-arabinose
MG1655	+	+	+
MG1655	+	+	-
MG1655	+	-	+
MG1655	+	-	-
MG1655	-	-	+
MG1655	-	-	-

The results are shown below, from which we could see 38.5 times of fluorescence/OD600 difference between transform group (without the expression of dCas9) and co-transform group(with the expression of dCas9) in RFP and 3.96 times in eGFP, indicating our transcription inhibition system has better effect on target site than lure sites, which matches our prediction.

Raw data: Transcription inhibition system verification(pIn-RTL3)

Targeting progression

We use our kinetic model to infer the difference between the off-target binding process of SpCas9 and xCas9. The improvement of on-target rate by directed evolution is derived from a more stable system after PAM recognition. Meanwhile, the energy penalty for incorrect matching obviously increases due to the variation of the interaction between DNA and Cas9-sgRNA complex. The free energy of the entire system is generally in a higher state during the targeting progression so that Cas9 is less likely to cleave the off-target site. This kinetics interpretation for xCas9 proves that directed evolution method can help reduce the off-target effects.

Mutagenesis library construction

The dCas9 mutant libraries are constructed using error-prone PCR.

To restrict the mutagenesis region to dCas9, we first amplified the dCas9 fragment and pdCas9 backbone through high-fidelity PCR. Homologous arms for Gibson assembly were added to both dCas9 and the backbone during PCR process. Then this dCas9 fragment was used as a template for error-prone PCR using error-prone PCR kit to obtain dCas9 randomly mutagenized library.

Then the randomly mutagenized library and the backbone are Gibson assembled. And theassembled library was transformed into DH5α for later plasimid library extraction, which was then transformed into MG1655 carrying pIN-RTr to verify the dCas9 mutant library.

According to our design, we do not need to sequence the plasmids immediately after acquiring the mutant library, but only to verify the success of our mutant library by flow cytometry analysis.

As is shown in the figure below, after mutating dCas9, we can see a significant change in its inhibition effect of gene circuits. Before mutation, the bacteria which was transformed initial pdCas9 plasmid showed almost no red fluorescent characterization. But after the mutation, red fluorescence characterization is very obvious, which means dCas9 with inhibiting effect of circuit has a significant change, and the dCas9 proteins are improved a lot. This also illustrates that the error - prone PCR is effective, and the our mutant library is built successfully.

Mutagenesis library screening

To select dCas9 mutants with higher on-target level and lower off-target level, the obtained dCas9 mutant library and the wild-type dCas9 were transformed by electroporation into MG1655 carrying pIN-RTL3, respectively.

We used flow cytometer to examine fluorescence intensity. L-arabinose was added to induce the expression of RFP and eGFP .

The results are shown below, from which we can see few of the proteins were optimal before mutation (i.e. on-target and not off-target), but the proportion of target proteins in the dCas9 protein library after mutation increased significantly, from 0.066% to 0.46%. At the same time, the proportion of proteins in the mutant dCas9 library that were neither on target nor off target also increased significantly, which may be caused by the inactivation of some mutant proteins. So this suggests that we have constructed a relatively ideal mutant library. And there were both positive and negative mutants, which also indicated that our previous work was effective and the design was reliable. Due to the limited time, the subsequent rounds of sorting have not been done yet, and we will continue to do it..