CN116004719A

CN116004719A - Method for reducing CRISPRCas9 off-target effect

Info

Publication number: CN116004719A
Application number: CN202211602646.8A
Authority: CN
Inventors: 徐亮; 刘彦; 王阳; 张秋龙; 王亮亮
Original assignee: Sun Yat Sen University
Current assignee: Sun Yat Sen University
Priority date: 2022-12-13
Filing date: 2022-12-13
Publication date: 2023-04-25

Abstract

The invention discloses a method for reducing CRISPRCas9 off-target effect, wherein sgRNA in a CRISPR/Cas9 system forms an RNA-DNA duplex recognition target DNA sequence through complementary pairing through a base with the length of 20bp of a Spacer part, when the Cas9 protein combined with the sgRNA searches NGG along a DNA double chain, the RNA-DNA forms a structure similar to R-loop, the pairing is started from 1 base, the activity of the sgRNA can be influenced by regulating the length of a Stem part, the appropriate length is selected according to different sequences, the targeting function of the sgRNA is still maintained, and compared with the sgRNA which is not modified, whether the off-target effect is improved or not is observed; according to the invention, a base complementary pairing structure is formed with a spacer part, a Hairpin structure is introduced at the 5' -end of the sgRNA, so that the Hairpin-sgRNA with the effect of reducing off-target is constructed, and a novel thought is provided for solving the ubiquitous off-target phenomenon of a CRISPR/Cas9 system by optimizing the number of base pairs of the stem part for any specific targeting sequence and finding the Hairpin-sgRNA sequence capable of effectively improving off-target on the premise of not influencing the targeting capability.

Description

Method for reducing CRISPRCas9 off-target effect

Technical Field

The invention relates to the technical field of gene therapy, in particular to a method for reducing CRISPRCas9 off-target effect.

Background

CRISPR-Cas systems are adaptive immune systems in bacteria and archaea and have been engineered as powerful gene editing tools. Class I CRISPR-Cas systems use a polyprotein complex to target nucleic acids, while class II systems can function using a single Cas protein. Because the system is simpler, work for gene editing with CRISPR-Cas systems is mainly focused on class II CRISPR systems. One can easily adapt it for various applications.

Competition between viruses and prokaryotes has led to Cas-related proteins possessing a wide variety of types in different bacteria. Each Cas protein has its unique properties (e.g., preference of nucleic acid sequences, different requirements for pre-spacer adjacent motifs (PAMs), different sizes of Cass proteins), and advantages and disadvantages when applied. Thus, identification and characterization of class II CRISPR systems is an area of intense research, the primary goal of which is to find Cas proteins with special functions or improved properties.

Although these nucleases can be used for gene editing in cells outside of their natural environment, they also have a significant off-target effect, resulting in an unexpected DNA break at sites that are not perfectly complementary to the spacer sequence. Therefore, improving the specificity of these nucleases is an important challenge, especially when the method is intended to be applied in the field of gene therapy, off-target effect is a major problem that must be solved.

Disclosure of Invention

The present invention is directed to a method for reducing CRISPRCas9 off-target effect, so as to solve the above-mentioned problems in the prior art.

In order to achieve the above purpose, the present invention provides the following technical solutions: a method of reducing CRISPRCas9 off-target effects, the method comprising the steps of:

(1) The sgRNA in the CRISPR/Cas9 system forms an RNA-DNA duplex recognition targeting DNA sequence through complementary pairing through a base with the length of 20bp of a Spacer part;

(2) When the Cas9 protein combined with sgRNA searches NGG along the DNA duplex, RNA-DNA will form a structure resembling "R-loop" and pair starting from base 1;

(2.1) pairing is successful until 20bp along the strand, and after complete complementary pairing, a stable RNA-DNA duplex is formed to activate HNH and RuvC nuclease domains of Cas 9;

(2.2) if there is a base mismatch in the middle, the RNA-DNA duplex will dissociate and the sgRNA-Cas9 protein complex will continue to find the next PAM site until the correct target pairing is found to be successful.

(3) The activity of the sgRNA can be influenced by regulating and controlling the part length of the Stem, and proper length is selected according to different sequences, so that the targeting function of the sgRNA is still maintained, and then compared with that of the non-modified sgRNA, whether the off-target effect is improved or not is observed;

(4) 200ng of the plasmid containing the EGFP gene is incubated with 50nM sgRNA and 50nM Cas9 enzyme in 1 XCas 9 bTfer for 30min at 37 ℃,1 mu L of protease K is added for further incubation for 10min to digest Cas9 protein, DNA loading buffer is added, the reaction solution is added into the prepared 1% agarose gel, electrophoresis is carried out in 1 XTAE buffer for 30min at a constant pressure of 140V at room temperature, and the plasmid is taken out and put into a gel imager to observe the plasmid cleavage;

(5) Twelve sgrnas of stem-4bp-stem-20bp were transcribed, and purity of the sgrnas was confirmed by a 6% polyacrylamide gel containing 7M urea after purification;

(6) Respectively quantifying the bands on the gel map, forming an intramolecular hairpin structure by extending at the 5' -end of the sgRNA, wherein the sgRNA and the targeting DNA form an RNA-DNA duplex when paired, and the sgRNA and the hairpin-sgRNA have a competition relationship;

(6.1) when the number of pairs is small, as the RNA-DNA duplex extends from the 20 th base of the sgRNA, the loop structure of the hairpin-sgRNA is opened, and eventually the 20bp RNA-DNA duplex is fully formed, activating the Cas9 enzyme, at which time the stem partial sequence becomes single-stranded RNA free outside the RNA-DNA duplex;

(6.2) if the number of pairs increases, the RNA-DNA duplex that has been formed is not stable enough to open the loop structure of the hairpin-sgRNA, then a 20bp complete RNA-DNA duplex will not be formed and Cas9 will not be activated.

(7) In cells, the off-target sequence of CRISPR/Cas9 is calculated through computer simulation, then the control group and the experimental group are compared through sequencing of potential off-target positions, whether the experimental system has off-target effect and the off-target frequency of each position can be determined, and in vitro, the corresponding target sequence is subjected to site-directed mutation through molecular cloning, so that plasmids with different target sequences can be obtained;

(8) Using a double-plasmid system, wherein one plasmid expresses sgRNA by taking U6 as a promoter, and the other plasmid expresses Cas9 protein by taking CMV as a promoter, and the two plasmids are simultaneously transferred according to a certain proportion;

(9) Cloning the sgRNA of the targeting related sequence and the modified hairpin-sgRNA into a plasmid pSLQ1651-sgTelomere (F+E) (Addgene, # 51024) by molecular cloning, and designing the stem-4bp-stem-12bp sgRNA with different pairing lengths according to the result of in vitro experiments for the hairpin-sgRNA;

(10) Two pairs of primers are selected to amplify target sequences, and the specificity of primer amplification under different temperature conditions is tested first;

(11) The experimental groups for respectively amplifying and transcribing different sgRNAs by using the EMX1-primer1 comprise an NC group, a PC group and a nine groups of samples of the sgEMX1-stem-5bp/6bp/7bp/8bp/10bp/11bp/12 bp;

(12) Screening out proper primers and annealing temperature at two positions with maximum off-target determined by a sequencing method, and respectively carrying out PCR amplification on gDNA of nine experimental groups;

(13) The recovered PCR product is used for T7E1 assay, the detection of a cutting strip in a target assay is analyzed, two sites targeting VEGFA are selected, and simultaneously, the PCR amplification research is carried out at the position with the highest off-target rate;

(14) By forming a base complementary pairing structure with the spacer portion, a Hairpin structure is introduced at the 5' -end of the sgRNA to construct a Hairpin-sgRNA with reduced off-target effects.

Preferably, in the step (1), the specific targeting key site is in a 20bp sequence, wherein each position may be any one of A, T, C, G bases.

Preferably, in step (2), if the 20bp base pair is to be fully complementary to the paired Cas9 protein to function, cas9 can be precisely targeted to the target site without causing off-target.

Preferably, in the step (6), if the number of pairs is extremely large, the number of initial base pairs of the RNA-DNA duplex is less than 5, and it is considered that the RNA-DNA duplex cannot form a stable RNA-DNA duplex, and thus it is considered that a mismatched sequence is encountered and randomly detached from the DNA strand, similarly to the sgRNA-stem-15bp/18bp/20bp.

Preferably, in the step (7), in vitro operations are performed with the target DNA, namely, plasma-WT, plasma-mt 18, plasma-mt 15, plasma-mt 12, plasma-mt 9, plasma-mt 6, plasma-mt 3, and the cleavage effect of different sgRNAs is tested.

Preferably, in the step (10), the annealing temperature is 68.0/67.4/66.4/64.9/63.1/61.6/60.6/60.0 ℃, and the DNA marker band sizes are 100bp,250bp,500bp,750bp,1000bp and 2000bp, respectively.

Compared with the prior art, the invention has the following beneficial effects:

1. according to the invention, a base complementary pairing structure is formed with a spacer part, a Hairpin structure is introduced at the 5' -end of the sgRNA, so that the Hairpin-sgRNA with the effect of reducing off-target is constructed, and a novel thought is provided for solving the ubiquitous off-target phenomenon of a CRISPR/Cas9 system by optimizing the number of base pairs of the stem part for any specific targeting sequence and finding the Hairpin-sgRNA sequence capable of effectively improving off-target on the premise of not influencing the targeting capability.

2. The scheme has the advantages of universality on any sequence, simpler design, no additional increase of system complexity, and pairing with the spacer part of the sgRNA through 3' -extension, so that the effect of reducing the off-target efficiency can be achieved, and the activity of the sgRNA is not influenced.

3. By designing an extended hairpin at the 5' -end of the sgRNA and partially modifying the secondary structure of the RNA in a spacer sequence (spacer), the obtained hairpin structure can improve the energy required by R-loop formation when the sgRNA-protein complex targets DNA, thereby improving the system specificity and achieving the effect of reducing off-target.

4. By adjusting the intensity of the secondary structure, the target position R-loop can still be formed normally, since RNA-DNA mismatch reduces the stability of the system, the increased secondary structure of sgRNA results in blocked R-loop formation at the off-target site, since R-loop formation is a critical process controlling the conformational transition of SpCas9 into active nuclease, which will block off-target site activating Cas9 nuclease activity and thus reduce off-target.

5. By accommodating the 5 '-end extension and forming an RNA-RNA duplex of the hairpin structure without interfering with the formation of the sgRNA-protein complex, the RNA hairpin generally follows Watson-Crick base pairing rules, whereas the 5' -end extension of the sgRNA requires only a stretch of sequence added on the basis of the original transcription, and does not add to the complexity of the system.

Drawings

FIG. 1 shows two ways of repairing the double strand breaks in cells according to the invention: non-homologous end joining and homologous recombination;

FIG. 2 is a schematic diagram of the structural design of hairpin sgRNAs of the invention;

FIG. 3 shows the sequence of the Hairpin-sgRNA-4 bp-20bp of the EGFP sequence targeting in the invention;

FIG. 4 is a gel electrophoresis of the stretch unpaired sgRNA cleaving plasmid of the present invention;

FIG. 5 shows the gel electrophoresis of (A) 6% modified acrylamide containing 7M urea, of the sgRNA of the present invention from stem-4bp to 20bp, (B) cleavage of plasmid and (C) cleavage of FAM-labeled 59bp double-stranded DNA;

FIG. 6 shows the quantitative results of (A) cleavage of plasmid and (B) cleavage of FAM-labeled 59bp double-stranded DNA according to the present invention;

FIG. 7 is a schematic representation of the structure formed by "R-loop" in the spacer binding of sgRNA of the present invention to target DNA;

FIG. 8 shows gel electrophoresis of sgRNAs and Cas9 cleaved plasmids of different designs of the present invention;

FIG. 9 is a schematic diagram of the T7E1 assay of the present invention for detecting mutations at specific sites in genomic DNA;

FIG. 10 shows the sequences of six gene loci and off-target loci according to the invention;

FIG. 11 is a sequence diagram of the cleavage site for expressing sgRNA in eukaryotic cells constructed by pSLQ1651-sgTelomere (F+E) of the present invention as a plasmid vector.

Detailed Description

The following description of the embodiments of the present invention will be made clearly and completely with reference to the accompanying drawings, in which it is apparent that the embodiments described are only some embodiments of the present invention, but not all embodiments. All other embodiments, which can be made by those skilled in the art based on the embodiments of the invention without making any inventive effort, are intended to be within the scope of the invention.

A method for reducing CRISPRCas9 off-target effect, wherein, the eukaryotic cell genome DNA extraction step is to wash cells with PBS, digest and centrifuge the cells with pancreatin to collect the cells at the bottom of a centrifuge tube, and the experiment adopts Wizard Genomic DNA Purification Kit of Promega company to extract gDNA of the cells according to the following steps:

(1) 600. Mu.L of nuclear lysis solution was added, and after mixing by pipetting, 3. Mu.L of RNase was added and incubated at 37℃for 15min, and then the solution was taken out and cooled to room temperature.

(2) After 200. Mu.L of the protein precipitation solution was added, the solution was vigorously shaken by a vortex shaker and placed on ice for 5min, at which time a white precipitate was observed to precipitate.

(3) Centrifuge 13000 Xg for 4min.

(4) The supernatant was carefully aspirated into a clean centrifuge tube and 600 μl isopropanol was added.

(5) Mixing the materials upside down.

(6) Centrifuge 13000 Xg for 1min.

(7) The supernatant was aspirated, and 600. Mu.L of 70% ethanol solution (at room temperature) was added.

(8) Centrifuge 13000 Xg for 1min.

(9) At this time, a small amount of milky precipitate (gDNA) was observed at the bottom of the tube, and the ethanol was carefully and completely aspirated off and left at room temperature for 15min, so that the ethanol on the surface of the remaining gDNA was volatilized clean.

(10) 100. Mu.L DEPC-H was added ₂ O,65℃was added for 60min to completely dissolve gDNA. After quantitative use, the mixture is placed in a refrigerator at the temperature of minus 20 ℃ for long-term storage

(11) After extracting gDNA, respectively making PCR reaction on target and off target by using correspondent primer, after recovering gel, making quantitative, taking 100ng PCR product to prepare 10 mu L T E1 reaction system, after the PCR product is annealed in NEB Buffer 2, adding 0.5 mu L T7 endonucleolytic I to make reaction at 37 deg.C for 15min, adding 0.5 mu L protease K to make reaction for 10min, adding 1% agarose gel and making electrophoresis for 30min.

sgRNA (single guide RNA) in the CRISPR/Cas9 system recognizes the targeting DNA sequence by complementary pairing of bases 20bp in length through the Spacer moiety to form an RNA-DNA duplex, thus the critical site for specific targeting is in this 20bp sequence. Wherein each position can be any of the A, T, C, G four bases, theoretically up to 1.1X10 of sequences to which CRISPR/Cas9 can target ¹² Personal (4) ²⁰ ). Whereas the human genome has about 3X 10 ⁹ bp base pairs, the system can target almost any desired site (limited by the sites actually targeted by Cas9 protein recognition PAM requiring NGG bases). When the NGG is searched along the DNA duplex by the sgRNA-bound Cas9 protein, the RNA-DNA will form a structure resembling an "R-loop" (fig. 1) and begin pairing from base 1. After successful pairing, the primer pair reaches 20bp along the strand, and after complete complementary pairing, stable RNA-DNA duplex is formed so as to activate HNH and RuvC nuclease domains of Cas 9; if there is a base mismatch in the middle, the RNA-DNA duplex will normally dissociate, and the sgRNA-Cas9 protein complex will continue to find the next PAM site until the correct target pairing is found to be successful. If the 20bp base pair were to be fully complementary to the paired Cas9 protein to function, cas9 can be precisely targeted to the target location without causing off-target. The actual situation is that after RNA-DNA duplex formation, individual bases appearIn case of unpaired, since the already formed duplex has a certain stability, it is possible that base pairing has continued along the strand direction when the hybridized duplex has not been dissociated, with the end result that the formed RNA-DNA duplex has individual mismatched base pairs in between, but the overall duplex stability is high enough to activate the Cas9 domain for function. At this time Cas9 will act at a non-target location, referred to as "off-target".

The experimental design of the Hairpin-sgRNA is shown in FIG. 2, and consists of a "Stem" part, a "GAAA-loop" part and a targeting sequence "Spacer" part. Wherein the sequence of the Spacer can be used for determining the base sequence according to different targeting positions, and the length of the Stem part is adjustable: the maximum can be 20bp, the minimum can be 4bp (the effective intramolecular hairpin structure can not be formed due to thermodynamic instability). The activity of the sgRNA can be influenced by regulating and controlling the part length of the Stem, and the appropriate length is selected according to different sequences, so that the targeting function of the sgRNA is still maintained. And then comparing with the sgRNA which is not modified, and observing whether the off-target effect is improved.

We select a segment of sequence (plasmid pEGFP-N1-FLAG map see appendix 6.4.7) "CCGGCAAGCTGCCCGTGCCCTGG" of targeting EGFP gene 152bp-714bp as target DNA, and design targeting sequence (PC group) of original sgRNA as 5'-GGGCACGGGCAGCUUGCCGG-3', and design thirteen sgRNAs with Stem length of 4bp, 5bp, 6bp, 7bp, 8bp, 9bp, 10bp, 11bp, 12bp, 15bp and 20bp (shown in figure 3). It is noted that, since the 5 '-end of the sgRNA must contain at least 3G bases by transcription with T7 RNA polymerase, we add a "-GGG-" sequence to the 5' -end of the sgRNA for consistency of the effect on experimental results. One problem with the introduction of additional G bases is that the number of pairs actually formed is likely to be greater than what we designed (G bases can base pair with C). For example, stem-5bp, stem-9bp, and Stem-12bp, since the spacer sequence has exactly one unpaired C base, the actual logarithm is 6bp, 10bp, and 12bp, respectively. In the research experiments we designed dense base sequences and therefore were difficult to avoid, which can be avoided in the actual design or taken into account the actual "G-C base pairing". For convenience we also named these sgrnas according to the design scheme. Meanwhile, considering that 5' -extended bases may affect the activity of sgrnas, we designed sequences in which the extended sequences do not pair with the spacer portion of the sgrnas, three sgrnas with base lengths of 10bp, 15bp and 20bp, respectively, were used as control groups (fig. 2).

First we explored whether 5' -end extension of the sgrnas of unpaired bases would affect the cleavage activity of Cas9 in the system. 200ng of the plasmid containing the EGFP gene was incubated with 50nM sgRNA, 50nM Cas9 enzyme in 1 XCas 9 bTfer for 30min at 37 ℃. Incubation was continued for 10min with 1 μl protease K added to digest Cas9 protein. After DNA loading buffer addition, the reaction solution was added to the prepared 1% agarose gel, and the gel was electrophoresed in 1×tae buffer at room temperature and constant pressure of 140V for 30min. Taking out and putting into a gel imager to observe the cutting condition of the plasmid. As shown in FIG. 4, NC-set plasmid in supercoiled state has the fastest electrophoresis rate, while plasmid cleaved by Cas9 into linear electrophoresis rate is slower, and the nicked plasmid with one strand broken has the slowest rate. We see that the PC group plasmid was completely cut into linearity, and its cleavage activity did not significantly affect as the 5' -end unpaired extension base increased to 20 nt. Slightly different from the 15bp and 20bp fragments, cleavage was seen to generate a partial nicked plasmid, indicating that 5' -extension of the excessive unpaired base does not significantly affect cleavage efficiency but does have a small effect on the cleavage function of Cas9. However, in general, the mere extension of the unpaired base at 5' -will not significantly affect the enzymatic activity. Therefore, the sgrnas of the hairpin structure designed by us can ignore the effect of the simple length extension on the Cas9 protein function.

The effect of hairpin-sgrnas on Cas 9's ability to cleave DNA was then investigated. We transcribed twelve sgRNAs of stem-4bp-stem-20bp and after purification we confirmed the purity of the sgRNAs by a 6% polyacrylamide gel containing 7M urea, as shown in FIG. 5A. The transcribed sgrnas can be seen to be of higher purity. We explored their ability to cleave plasmid DNA (FIG. 5B) and 59bp long short-chain dsDNA (FIG. 5C), respectively. For cleavage of plasmid DNA, it can be seen that with increasing logarithm, the plasmid is cleaved to a lesser extent, and no cleavage band has been observed at all by the time of sgRNA-stem-12 bp. Indicating that the sgrnas at this time have lost their ability to target DNA. While for shorter-chain dsDNA that is more difficult to cleave, the cleavage capacity drops faster, and almost no cleavage product is obtained by the time the sgRNA-stem-10bp is reached (59 bp FAM-dsDNA band followed by 34bp base cleavage product).

The bands on the gel pattern were quantified separately and the results are shown in FIG. 6. We can more intuitively see that with increasing stem length, sgrnas gradually lose their ability to target DNA.

Table 4-1 shows the theoretical Gibbs free energy of base pairing (IDT OligoAnalyzer, HAIRPIN, target type: RNA; oligo Conc: 0.25. Mu.M; na) as the number of stem portion pairs of Hairpin-sgRNA increases ⁺ Conc:50 mM). We did not predict the theoretical gibbs free energy of the RNA-DNA duplex because this step has the Cas9 protease involved in forming "R-loop" and therefore its stability is much more stable than a simple double-stranded RNA-DNA duplex of the same sequence.

Although different targeting sequences will result in a difference in the number of extensions, the thermodynamic stability of base pairing ultimately plays a decisive role. We previously talk that by extending the 5' -end of the sgRNA to form an intramolecular hairpin structure, the formation of RNA-DNA duplex when the sgRNA is paired with the targeting DNA is in a competing relationship with the hairpin-sgRNA (FIG. 7). When the pairing number is small, as the RNA-DNA duplex extends from 20 bases of the sgRNA, the loop structure of the hairpin-sgRNA must be opened, and finally, the 20bp RNA-DNA duplex is completely formed, and the Cas9 enzyme activity is activated, and the stem partial sequence becomes single-stranded RNA to be free outside the RNA-DNA duplex. In contrast, if the number of pairs increases, the RNA-DNA duplex that has been formed is not stable enough to open the loop structure of the hairpin-sgRNA, then a 20bp complete RNA-DNA duplex will not be formed and Cas9 will not be activated. There are, of course, more extreme cases where the number of pairs is extremely large, like sgRNA-stem-15bp/18bp/20bp. Since the number of initial base pairing of RNA-DNA duplex is less than 5, it is considered that stable RNA-DNA duplex is hardly formed, and thus it is considered that a mismatched sequence is encountered and randomly detached from the DNA strand.

TABLE 4-1 Gibbs free energy prediction of the paired number of Hairpin-sgRNA formation

Table 4-1 The Gibbs prediction forming the base pair number of hairpin-sgRNA

/>

Next we studied in vitro off-target experiments with part of the Hairpin-sgRNA. In the cell, the off-target sequence of CRISPR/Cas9 is calculated through computer simulation, and the existence of off-target effect and the off-target frequency of each site of an experimental system can be determined by comparing a control group with an experimental group through sequencing of potential off-target positions. In vitro, the method is relatively simple, and plasmids with different target sequences can be obtained by carrying out site-directed mutagenesis on corresponding target sequences through molecular cloning. And taking the same as target DNA, respectively, plasma-WT, plasma-mt 18, plasma-mt 15, plasma-mt 12, plasma-mt 9, plasma-mt 6, plasma-mt 3, and testing the cutting effect of different sgRNAs (Table 4-2).

In addition to PC group, experimental group we selected RNA group (sgRNA-stem-4 bp-10bp, cleavage activity > 85%) with no significant effect on cleavage plasmid activity, and in view of no significant difference in the experimental results of sgRNA with log less than 8bp, we selected gel electrophoresis pattern with representative sgRNA-stem-8bp as representative for analysis. First, we cut with unmodified sgrnas, plasmids that were perfectly matched and incompletely complementarily matched with one base mutation at a different position, respectively. Experimental results show that, similar to the nearly 100% cleavage capacity of the fully complementary pairing, the sgRNA-Cas9 protein complex is not recognized for single point mutated targeting sequences, i.e. for these off-target sequences, non-engineered sgRNA-targeted sequences will cause non-specific cleavage of Cas9, resulting in off-target phenomena (fig. 8A). In the case of the hairpin-sgRNA, however, when the number of pairs is small, for example, the stem portion is.ltoreq.8 bp (the different sequences differ slightly due to the different stabilities only for the sequences of this experiment). Because of the too small number of pairs, RNA-DNA dTplex already forms a number of pairs exceeding 12bp when competing with the loop portion of the hairpin-sgRNA. At this point, pairing cannot continue even when a single non-complementary paired base is encountered until the loop portion of the hairpin-sgRNA is completely replaced to form the complete 20bp RNA-DNA dTplex. The experimental results show that the sequence of single base mutation can not be identified similarly to the PC group sgRNA, and the off-target phenomenon is caused (FIG. 8B).

Whereas for sgRNA-stem-9bp (FIG. 8C) and sgRNA-stem-10bp (FIG. 8D), different properties were exhibited. The partially mutated sequences mt-9 and mt-18 can be identified, thereby reducing off-target efficiency. Not so consistent with our guess, we began to think that for sgrnas that improved off-target, the closer the target sequence mutation was to position 20, the more significant the improvement should be. Since the pairing of RNA-DNA duplex is initiated from position 20 to position 1, the previous literature reported that the sequence of about 10 to 12nt from position 20 to 10 of sgRNA belongs to the "seed region", which is extremely critical for specific recognition and binding of DNA. If the partial sequence cannot form better pairing, the formation of the sgRNA-Cas9-DNA ternary complex is affected, and then the Cas9 is affected to play a role. In practice, however, except that mt-18 is closer to the 20 bases, mt-15 and mt-12 did not significantly improve off-target, but mt-9 also improved off-target (the ability to cleave the plasmid DNA was weaker).

By analysis, we make the following explanation: base complementary pairing of nucleic acids is a thermodynamic process, that is, if unpaired G/C bases and C/G bases or A/T bases and T/A bases are close to each other, then complementary base pairs tend to form by hydrogen bonding. Because of the randomness of the process, base pairs that were successfully paired first are intended to remain stable and must rely on nearby bases to continue to form paired base pairs to reduce overall energy. If the pairing is successful, the pairing can be completed rapidly along the chain like a zipper; conversely, if the nearby bases do not pair, then a small piece of the paired base that has formed will dissociate due to random perturbation between the molecules, thus beginning to repeat the previous step. Until a small portion of the bases pair successfully and extend along the strand to a pairing length that is sufficiently stable at that temperature, at which point the successfully paired strand will no longer dissociate, but rather exists in the least energetic paired form of a double strand.

That is, the sgRNA-Cas9 protein complex, after searching for PAM sequences on the DNA strand, is split by the protein-assisted DNA duplex and unpaired by the sequence of the spacer portion of the sgRNA. In practice, it will not only begin pairing to position 1 at bases near position 20. Instead, the targeting sequence of the sgRNA will begin to extend complementary pairing along both ends of the strand as some portion of the base pair randomly on the DNA portion in the single stranded state until 20nt of base pairs are completed. Further, since the 5' -sgRNA has a larger degree of freedom in part, the partial sequence may preferentially pair with each other, and thus, the partial sequence may be paired with a 20-base portion in the middle, thereby causing a phenomenon of "zipping" together in the middle. Complete pairing is successful in activating Cas9 nuclease activity. Where there is an unpaired middle, this will result in a cessation of pairing extension, while the portion of the pairing that has been complemented at both ends will not be sufficient for its alone stability, and will eventually tend to dissociate.

Returning to the hairpin-sgRNA design of our design, if this pairing pattern is followed, then the partial improvement of the plasmids mt-9 and mt-18 by the sgRNA-stem-9bp and the sgRNA-stem-10bp can be explained. For mt-18, too close to the 20-base position can result in unpaired portions of the sgRNA-Cas9 protein complex being encountered upon 2bp base pairing after the PAM sequence on the DNA strand is searched, where the 5' -portion of the sgRNA cannot aid in pairing due to the presence of a more stable intramolecular hairpin structure. And thus, a stable RNA-DNA duplex cannot be formed, so that the sgRNA-Cas9 protein complex breaks away from the DNA strand to find the next target region. The mutated DNA sequence is not cut effectively, improving off-target. When the mutation position is close to 20 bits, the sequence of the hairpin structure formed by the hairpin-sgRNA can complete the pairing process together with the right part, and when the two parts are successfully paired, the hairpin structure formed by the hairpin-sgRNA is partially opened due to the action of chain exchange, and finally, the hairpin-structure is completely matched; when the mutation position is close to the 1 position, the hairpin can be completely opened by pairing the 20 positions. In the case of single point mutant plasmid mt-9, the mutation position is located in the middle part of the 20nt mating sequence. At this time, the portion paired from the 20-position base position is not stable, and the hairpin structure of the hairpin-sgRNA is not opened, so that the complete complementary paired RNA-DNA duplex cannot be formed.

Overall, in vitro experiments demonstrated that the targeting ability of the sgrnas of the hairpin structure decreased with increasing logarithm of the stem portion. However, the targeting ability in the intermediate sequence is not obviously reduced, and the structural design for distinguishing single-base mutant DNA can be realized. We therefore want to further apply this structural design to cell experiments.

It must be noted at first that in vitro experiments differ significantly from cellular experiments, albeit both are CRISPR/Cas9 systems. For example, in an in vitro experimental system, the components are relatively simple and controllable, and in one reaction system, only three main components of simple sgRNA, cas9 protein and dsDNA are included, wherein the sgRNA and Cas9 protein are added in a ratio of 1:1 after being quantified. dsDNA is also simpler, even if plasmid DNA is used as target DNA, its sequence is no more than 10kb (the plasmid used in this experiment is approximately 4.7 kb). For one sgRNA, there is little second target in the system. The target sequence is subjected to site-directed mutagenesis in the off-target experiment. In general, in vitro experiments are relatively simple and controllable.

TABLE 4-2 target DNA and point mutation sequences thereof

Table 4-2 Target DNA and its point mutation sequence

For cell experiments we used a two plasmid system: one plasmid expressed sgRNA with U6 as promoter (plasmid map 6.4.6) and the other plasmid expressed Cas9 protein with CMV as promoter (plasmid map 6.4.10). Simultaneously transferring two plasmids according to a certain proportion, and in terms of space time, since the expression of the Cas9 protein needs two steps of transcription and translation, the Cas9 protein can appear later than sgRNA. Meanwhile, if the two are not in the same plasmid and the promoters are different, the sgRNA expressed in the cells and the translated Cas9 protein can be certainly different in quantity, and cannot be 1:1. The most complex is that human genomic DNA sequences are up to 30 hundred megabase pairs, while targeting sequences are found by means of only 20bp spacer sequences. Its off-target sequence is much more complex. So far one can still only find out theoretically possible off-target sequences by computer, and the number of mismatched bases can be as high as four to five. Experiments show that, where some targeting sequences are identical but PAM sequences are not present, off-target phenomena can still be detected, which indicates that the mechanism of action of Cas9 in vivo is not exactly the same as in vitro (experiments in vitro indicate that, without PAM sequences, even if the spacer portion 20bp sequence of sgRNA is exactly the same as the DNA portion, cas9 enzyme is completely ineffective).

Most importantly, the in vivo detection of off-target is quite different from in vitro experiments in that in cells, as shown in fig. 1, cas9 enzyme cleaves genomic DNA resulting in Double Strand Breaks (DSBs), cells repair broken DNA by an intrinsic mechanism, and the pathways are largely divided into two: non-homologous end joining (Nonhomologous end joining, NHEJ) and homologous recombination (Homology directed repair, HDR). Of these, NHEJ is the most common method, and this process directly links the broken portions of DNA by related proteins, and also causes deletion, insertion, and substitution of bases in the broken portions of DNA. The HDR process is complicated, and by transferring ssDNA or plasmid DNA with homologous sequences at two ends of DNA fragmentation as a template for homologous repair, the enzyme associated with the pathway will completely insert the sequence between the two homology arms of the donor DNA into the newly repaired DNA, and by using this strategy one can insert the sequence of interest in the part of the sequence to be studied, i.e. achieve "knock-in" of the gene.

The DNA repaired by both pathways has corresponding detection methods: for homologous recombination, it is common practice to insert a small Duan Te-specific restriction enzyme-recognized sequence into the donor DNA, design primers upstream and downstream of the sequence in subsequent experiments, amplify the sequence by PCR, and then cleave the amplified fragment with a restriction enzyme, and for sequences that have not been edited, the cleavage is not possible due to the lack of the recognition site for the enzyme. Finally, the efficiency of gene editing (knock-in) can be assessed by the percentage of cleavage.

The detection method of non-homologous end joining is somewhat complicated (as shown in FIG. 9). Cells are collected and their genomic DNA is extracted after the transfection experiment (step 1) is completed (step 2), and then the corresponding sites are amplified by PCR method by designing specific primers (step 3), including unedited original sequence and DNA cleavage after Cas9 cleavage, cell repair introducing base deletions, insertions and substitutions (yellow part). And then carrying out heating denaturation and rapid cooling double-strand annealing on the amplified PCR product, wherein double strands are randomly paired together in a base complementary pairing mode, and partial products consist of an original sequence and a mutant sequence. That is, the partial complementary pairing of the intermediate portion of the strand sequence is partially incomplete (step 4). At this time, T7E1 enzyme is added, which specifically recognizes and cleaves the portion of the double strand that is not fully complementary (step 5). By agarose gel electrophoresis, a cut band can be observed, and whether or not a band is present or not and the intensity of the band can be observed to determine whether or not Cas9 has DSB and efficiency of action at the sequence (step 6).

After determining the experimental method and the detection means, we find some reported positions which are relatively easy to off-target by consulting the literature, and the sequences related to the target (WT) and the off-target (OT) are shown in FIG. 10. In the literature, potential off-target sequences are found out through comparison of human gDNA big data, amplified and sequenced after experiments, and the actual off-target efficiency of the potential off-target sequences can be confirmed by comparison with a control group. We selected the EMX1 gene, FANCF gene, VEGFA gene (site 1/site 2), RUNX1 gene, ZSON 2 gene and the seven sequences with the highest off-target rates for research.

By molecular cloning, we cloned both the sgrnas targeting the relevant sequences and the engineered hairpin-sgrnas into the plasmid pSLQ1651-sgTelomere (f+e) (adedge, # 51024), part of the sequence is shown in fig. 11. The plasmid takes U6 as a promoter, and can efficiently express sgRNA in eukaryotic cells. Meanwhile, for the hairpin-sgRNA, according to the results of in vitro experiments, we designed stem-4bp-stem-12bp sgRNAs with different pairing lengths. Also, since the 5' -end of the plasmid inserts the sgRNA sequence through BstX I cleavage site, there is an extra G base during transcription of the sgRNA, so that there is a partial sequence just matching the G base and thus increasing the length of stem, we have no repeated transcription of RNA that actually contains the same number of paired bases in order to avoid repetition. Specific hairpin-sgRNA sequences are shown in tables 4-3.

TABLE 4-3 design of the Hairpin-sgRNA Structure and the sequences thereof

Table 4-3 Hairpin-sgRNA structure design and sequence

/>

/>

/>

We chose that the transfected cells were HEK293T, the specific experimental procedure was as follows: spread 15 ten thousand cells to six well plates, and transfect 2.5 μg of plasmid pSLQ-1651-sgRNA encoding sgRNA and 2.5 μg of plasmid pEGFP-N1-Cas9 encoding Cas9 protein with lipo3000 liposome transfection reagent. After 6h medium was supplemented to a final volume of 1mL. After 24h of incubation, red fluorescence (mcherry carried by the plasmid encoding the sgRNA) and green fluorescence (EGFP carried by the plasmid encoding the Cas9 protein) were observed under a fluorescence inverted microscope, indicating that both the sgRNA and Cas9 proteins were expressed normally. Cells were collected 48h after transfection. Genomic DNA of the cells was extracted using Wizard Genomic DNA Purification Kit from Promega company, and then the corresponding sites were amplified by PCR method by designing specific primers. And then carrying out heating denaturation and rapid cooling double-strand annealing treatment on the amplified PCR product, wherein double strands are randomly paired together in a base complementary pairing mode, and partial products consist of an original sequence and a mutant sequence. That is, the sequence of the intermediate portion of the strand will have a partially incompletely complementary pair. Then adding T7E1 enzyme, observing the cut band and the band strength through agarose gel electrophoresis, and determining whether Cas9 has the effect and the efficiency of DSB at the sequence.

The most important step in this experiment is the specific amplification of a specific site. Genomic DNA is known to contain up to 30 hundred megabase pairs, and it is necessary to amplify it by a pair of primers (about 20nt long upstream and downstream primers), and therefore primer selection is of paramount importance. The poor design of the primers can result in the amplified product being so impure that multiple amplified bands appear, and even worse, no apparent band of interest can be observed after the PCR reaction. By using the Primer-BLAST software of NCBI website (https:// www.ncbi.nlm.nih.gov/Tools/Primer-BLAST/index. Cgilink_LOC=blast Home), the convenience of designing primers can be greatly improved. By setting the relevant parameters, non-specific binding of the primer can be avoided as much as possible, thereby amplifying fragments of a specific length comprising the band of interest. It should be noted that although the website will give the optimal primer design according to the parameters we set. However, in practical application, non-specific bands can be amplified, some of the bands can be increased in purity by increasing the annealing temperature, some of the bands can not be amplified under any optimized condition, and only the primers can be redesigned. After PCR amplification, the target amplified bands are collected by a gel recovery method, so that the purity of the target bands cut in the subsequent T7E1 test is ensured. The upstream and downstream primers for amplifying the relevant target sequences and the amplified band sizes are shown in tables 4-4.

Tables 4-4 amplified genes were designed at the target and off-target site primer design and PCR product size.

Figure 4-4 Primer design and PCR product size amplifing each gene at on or off-target site.

/>

/>

For amplifying EMX1 sequences, two pairs of primers are selected to amplify target sequences, and firstly, the specificity of primer amplification under different temperature conditions is explored. The annealing temperatures were 68.0/67.4/66.4/64.9/63.1/61.6/60.6/60.0℃respectively (unless otherwise indicated hereinbelow, we all refer to the optimized primers annealing temperatures being between 60-68 ℃). The DNA marker bands were 100bp,250bp,500bp,750bp,1000bp and 2000bp (the same applies below). It can be seen that the specificity of the two pairs of primers is good, and the yield of EMX1-primer1 is higher than that of EMX1-primer 2. The temperature does not have a great influence on the EMX1-primer2, and the yield of EMX1-primer1 decreases with increasing temperature. Comprehensively considering that the amplification at a lower temperature is selected on the premise of ensuring the purity so as to improve the yield of the PCR product.

The experimental groups for transcription of different sgrnas were amplified with EMX1-primer1, respectively, including NC group (sgrnas not targeting any gene), PC group (sgEMX 1) and sgEMX1-stem-5bp/6bp/7bp/8bp/10bp/11bp/12bp, for a total of nine groups of samples, the maximum of two positions of off-target as determined by sequencing. After the proper primer and annealing temperature are screened out, the gDNA of nine experimental groups is respectively subjected to PCR amplification, the recovered PCR products are used for T7E1 assay, the detection of a cutting strip in a target assay is analyzed, two sites targeting VEGFA are selected, and simultaneously, the PCR amplification research is carried out at the position with the highest off-target rate.

Tables 4-5 summary of results of six Gene loci on target and off target experiments

Table 4-5 Summary of on-target and off-target experimental results of six gene loci

By forming a base complementary pairing structure with the spacer moiety, we constructed a Hairpin structure at the 5' -end of the sgRNA to construct a Hairpin-sgRNA with reduced off-target effects. We found, for any particular targeting sequence, a Hairpin-sgRNA sequence that can effectively improve off-target without affecting the targeting ability by optimizing the number of stem partial pairing bases. This provides a new idea for the CRISPR/Cas9 system to address the ubiquitous off-target phenomenon, summarized in tables 4-5.

Although embodiments of the present invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made therein without departing from the principles and spirit of the invention, the scope of which is defined in the appended claims and their equivalents.

Claims

1. A method of reducing the CRISPRCas9 off-target effect comprising the steps of:

2. The method of reducing the CRISPRCas9 off-target effect of claim 1, wherein: in the step (1), the specific targeted key part is in a 20bp sequence, wherein each position can be any one of A, T, C, G bases.

3. The method of reducing the CRISPRCas9 off-target effect of claim 1, wherein: in step (2), cas9 can be precisely targeted to the target site without causing off-target if the 20bp base pair is to be fully complementary to the paired Cas9 protein to function.

4. The method of reducing the CRISPRCas9 off-target effect of claim 1, wherein: in the above step (6), if the number of pairs is extremely large, the number of initial base pairing of the similar sgRNA-stem-15bp/18bp/20bp is less than 5, and it is considered that the RNA-DNA duplex cannot be formed stably, and thus it is considered that a mismatched sequence is encountered and randomly detached from the DNA strand.

5. The method of reducing the CRISPRCas9 off-target effect of claim 1, wherein: in the step (7), in vitro operations are respectively performed on the target DNA, namely, the plasma id-WT, the plasma id-mt18, the plasma id-mt15, the plasma id-mt12, the plasma id-mt9, the plasma id-mt6 and the plasma id-mt3, and the cutting effect of different sgRNAs is tested.

6. The method of reducing the CRISPRCas9 off-target effect of claim 1, wherein: in the step (10), the annealing temperature is 68.0/67.4/66.4/64.9/63.1/61.6/60.6/60.0 ℃ respectively, and the DNA marker band sizes are 100bp,250bp,500bp,750bp,1000bp and 2000bp respectively.