Zhang Lab's CRISPR Frequently Asked Questions
Dr. Le Cong, the first author of one of the seminal articles from Dr. Feng Zhang's Lab ( Science , February 2013 ), has assembled a list of FAQs about using the lab's CRISPR technology.
CRISPR Design FAQs
Should I use wildtype or double nickase for my CRISPR genome engineering experiments?
When assessing which nickase type to use for your CRISPR genome engineering experiments, consider that wildtype Cas9 with optimized chimeric gRNA has high efficiency but has been shown to have off-target effects. 'Double nickase' is a new system, developed by the Zhang lab, which has comparable efficiency to the optimized chimeric design but with better accuracy (in other words, lower off-target effect).
The double nickase system is based on the Cas9 D10A nickase described in Figure 4 of the Cong, et. al, 2013 Science paper. For example, if you want to use double nickase, you could express two spacers and use PX335 to express the Cas9n (nickase).
The concept of the double nickase system is that you can express two different chimeric gRNAs with the Cas9 nickase which will together introduce cleavage of the target site with efficiency similar to using a single chimeric gRNA. At the same time, the off-target effects are reduced because the Cas9 nickase doesn't have the ability to induce double-stranded breaks like the wildtype Cas9 does. There are a few references for the double nickase system, including one recently from the Zhang group.
Many protocols have been published with thorough step-wise instructions on designing and using Cas9-based genome engineering tools to perform different types of genome targeting. They are very good source of references. For genome editing, the Zhang lab has published this Nature Protocol article.
Should I add the PAM sequence to the oligo when cloning my target spacers into the PX330 vector backbone?
There is no need to add the NGG PAM sequence. Because the 'NGG' of the PAM is used to select your genomic target, you need to make sure the NGG immediately follows your target on the genome but NOT on the oligo.
The actual oligo you should order is:
You need to add the CACC and AAAC cloning overhang.
When designing oligos for cloning my target sequence into a backbone that uses the human U6 promoter to drive expression, is it necessary to add a G nucleotide to the start of my target sequence?
The human U6 promoter prefers a 'G' at the transcription start site to have high expression, so adding this G could help with expression, though it is possible for the plasmid to still express without the G. Because the G is only one base, the Zhang lab usually adds it when they order the oligo. If your spacer sequence starts with a 'G', you naturally have one and do not need to add an additional 'G'.
Where are double strand breaks (DSBs) induced, compared to where the target site sequence (protospacer+PAM) is located? If there are other PAMs in the region, will they also be targeted?
The Cas9 cuts 3-4bp upstream of the PAM sequence. There can be some off-target DSBs using wildtype Cas9. The degree of off-target effects depends on a number of factors, including: how closely homologous the off-target sites are compared to the on-target site, the specific site sequence, and the concentration of Cas9 and guide RNA (gRNA). These considerations only matter if the PAM sequence is immediately adjacent to the nearly-homologous target sites. The mere presence of additional PAM sequences should not be sufficient to generate off-target DSBs; there needs to be extensive homology of the protospacer followed by PAM.
Click here for a target selection tool that should be helpful in minimizing or testing off-target DSB.
Is it possible to target a single allele when the target sequence is present in both alleles?
It should be possible to target a single allele when the target sequence is present in both alleles. The Zhang lab has found that if a CRISPR is used to engineer a locus, isolate single cell colonies, then genotype them, it is possible to find both single-allelic and bi-allelic cells. Single-allelic cells usually make up the majority in culture unless the targeting efficiency is very high.
How should I design experiments for multiplexed targeting with Zhang lab plasmids?
For multiplexing CRISPR to target multiple genome loci, the most efficient and easiest way is to co-transfect several plasmid together, with each plasmid having a targeting spacer cloned into the backbone (pX330 or PX335, depending if you want to use wildtype cas9 or double nickase).
For example, if you want to target two genomic loci with wildtype Cas9, clone two spacers for each locus into PX330, then co-transfect or delivery the two plasmids together into your cells. If you want to use double nickase, you need to have four spacers co-expressed and using PX335 to express the Cas9n (nickase).
Homologous Recombination (HR) FAQs
Can you summarize HR template vector design guidelines?
A few notes below are considerations for designing HR donor.
Generally if off-target (non-specific) cleavage is not a very big concern, you could go ahead with PX330 plasmid, clone in your target guides, test run them and then select the best guide to co-transfect in HR donor. Alternatively, a new system described recently is the double nickase system, with PX335 vector, you need to find two guides for each cleavage site, but this system probably have better specificity. A good reference is this Cell paper: http://www.cell.com/abstract/S0092-8674(13)01015-5
As for target selection, we usually pick 3-6 guides around the region to find the most efficient guide (most guides would work, but they sometimes have different efficiency). And it is best if the cut site is as close to the junction of the homology arm as possible. Definitely should be less than 100bp away, ideally less than 10bp away.
For small (<50bp) changes, you should consider using ssDNA oligo as donor template, it's usually more efficient than plasmid donor. Typical design is to have 50-80bp of homology arms on each side flanking the change you'd like to introduce. We use Ultramer oligo from IDT (non-PAGE purified oligo is fine based on new data). For general transfection, you could use 1ul of 10uM oligo stock for each well on 24-well plate, co-deliver into the cells with the Cas9/sgRNAs vector (like PX330).
For large changes (>100bp insertions or deletions), we typically use plasmid donor, with two homology arms on each side flanking your desired insertion or mutation. Each arms is around 800bp. For transfection in cells like HEK, we use around 400ng of a typical size (~ 5kb) donor plasmid with Cas9/sgRNAs vector, for one well in 24-well plate. A good example/reference is this Cell paper: http://www.cell.com/abstract/S0092-8674(13)01016-7
In the case where you have intact CRISPR protospacer target in the HR template, you usually would need to make mutations of the HR template to avoid this donor plasmid being degraded by Cas9 in cells. One good way is to mutate the PAM 'NGG' in the HR template, like change it to 'NGT', or 'NGA', or 'NGC' (if it's within coding region make sure it's a silent mutation).
Is a protospacer followed by the PAM (my target sequence) in my HR template an issue?
Yes. If you have intact protospacer+PAM sequence within the HR template, it will likely be targeted and degraded by the Cas9. It is possible to make a silent mutation or avoid putting in the full target site in the donor. For example, choosing target sites that span the knock-in gene.
For making mutations, one good way is to mutate the PAM 'NGG' sequence in the HR template. Since the PAM sequence is required for successful targeting and cleavage, changing it to'NGT' or 'NGC', in addition to mutations in the spacer itself, will protect the HR template from the Cas9. Again, if it's within coding region make sure it's a silent mutation.
When attempting to use the CRISPR/Cas9 system to create specific mutations or insertions by Homologous Recombination (HR), what is the length of each homology arm needed? How should the double strand break (DSB) be positioned relative to where the homology ends?
The Zhang lab has always used homology arms that are less than 15bp away from the double strand break site. Longer distances will also work but the efficiency will be lower, although this might still be okay if you are introducing a selection marker gene. The homology arms should be no more than 100bp away from the DSB, ideally less than 10bp away if possible.
For introducing small mutations (<50bp) or a single-point mutation, the best HR template for transfection is probably a single stranded DNA (ssDNA) oligo which usually work better than plasmids. For ssDNA oligo design, the Zhang lab typically uses around 100-150bp total homology. The mutation is introduced in the middle, giving 50-75bp homology arms. The Zhang lab typically use PAGE purified long oligos. For large changes (>100bp insertions or deletions), the Zhang lab typically uses a plasmid donor, with two homology arms on each side flanking your desired insertion or mutation. Each arm is around 800bp.
For transfection in cells like HEK, the Zhang lab uses around 400ng of a typical size (~5kb) donor plasmid with Cas9/sgRNAs vector, for a one well in a 24-well plate. A good reference is this paper in Cell.
What is the maximum amount of DNA that can be inserted into the genome using CRISPR/Cas for Homologous Recombination (HR)? How long should the homology arms be for efficient recombination?
The most we've tried to insert so far has been 1kb. We used homology arms that were 800bp long.
Any reference for using ssODN as HR Template?
You can take a look at this paper (Soldner et al. , Cell. 2011) which uses ssODN with shorter homology arms.
At the Bench FAQs
How can I quickly check if the cloning of my oligo into a backbone likePX330worked?
You can quickly check which of your colonies carry the correct insert (gRNA) in pX330/pX335 with BbsI/AgeI double digestion. As a successful insertion will destroy the BbsI sites, a double digest should discriminate between positive and negative clones. Clones with insertion will show only linearized plasmid of ~8.5 kb (only AgeI will be able to cut). Clones without insertion will show a ~1kb and ~7.5kb fragment (both BbsI and AgeI will be able to cut).
After the introduction of a mutation into the genome, how can cells with that mutation be selected/screened?
Before starting your experiment, consider co-transfecting with GFP. This allows you to sort for GFP-positive cells and to enrich for those cells that were positively transfected. Alternatively, you can use a selection marker to select transfected cells (for example, plasmid with a puromycin resistance cassette, such as PX459). After you co-transfect the CRISPR/Cas system with your homologous recombination (HR) template, you could then:
Confirm your HR by doing Restriction Fragment Length Polymorphism (RFLP) (see Figure 4 of the Cong, et. al, 2013 Science paper).
If you detect positive HR, isolate single-cell colonies, grow them up, then perform individual genotyping (using Sanger sequencing, for example) on each colony in order to screen for positive ones.
- Or -
If your HR template has a selection marker such as puromycin, you can (also) select for the positive colonies by puromycin selection. You could then confirm this purification by performing a genotyping assay (such as Sanger sequencing).
Click here for a useful reference.
I have used 200ng to 500ng of 293FT genomic DNA as template after performing gDNA purification and have tried conditions with 5% DMSO or without 5% DMSO. My genomic PCR didn’t work, do you have any suggestions?
The Zhang lab recommends using the Epicentre QuickExtract solution to perform this step to extract genomic DNA from cells. The Zhang lab uses around 50ul of the extraction solution, runs the protocol, then uses around 2-4ul in each 50ul PCR reaction. The gDNA extracted in this way is usually more concentrated than using other kits like the Qiagen kit, so its better for PCR and surveyor.
I used DNA polymerase Takara Ex Taq™ DNA Polymerase for my genomic PCR, but couldn't amplify the EMX1 gene using same primer you used in the Science paper (Cong et al. , Science 2013) do you have any suggestions?
To date, I have never used the Ex Taq. In our hands, Herculase II Fusion polymerase or Kapa Hifi Polymerase work very well. Some groups have successfully used Herc II or Kapa Hifi in PCR. Maybe this Takara enzyme is not very robust in this case for EMX1.
Since the publication of our paper, we have two new optimized primers that may work better than the published ones, so if/when you use HercII or Kapa Hifi the reaction still does NOT work, you can try these new primers:
- EMX1-Forward: CCATCCCCTTCTGTGAATGT
- EMX1-Reverse: GGAGATTGGAGACACGGAGA
The reason for the 30nt region on the upper is because the top figure shows the construct from our 'Split RNA' design, where the pre-processing form of crRNA is expressed and a tracrRNA (the helper RNA that facilitate the processing of the crRNA) is also expressed from a separate promoter. In the pre-processing form, the crRNA contains 30nt of the target spacer, but then when expressed in cell and with Cas9 and tracrRNA, it will be processed (i.e., cleaved) into a shorter form that contains 20bp of the target spacer, or in another words, the first 10bp will be cut off. That's why in the lower figure with PX330 and PX335, where the construct is from our 'Chimeric RNA' design (meaning the mature processed form of crRNA is fused chimerically to a mature tracrRNA), the spacer used is only 20bp. Please refer to Cong et al. , Science 2013 for more details.