Unless you isolate 100% homozygous clones (where all alleles of the target gene have been edited) you will always detect expression by WB from unedited or heterozygous cells.

Dear Sehrish,

CRISPR create double stranded breaks which are then repaired. The reparation process has a chance of changing your sequence in a way that a protein retain its activity  (e.g. insertion of three nucleotides which does not affect protein stability) whereas preventing Cas9 from cleaving it again (since your target sequence has been changed). For this reasons it is good to isolate homozygous clones to get 100% KO.

However, if you are not isolating and doing the KO in a cell population, 3 active gRNAs should result in more than 99% cells having KO in both alleles (sometimes it is not possible to detect protein on a WB even after KO with one gRNA in a cell population). In your case you have 50% KO which looks like either your gRNAs has really low activity (which is unlikely) or your gene is essential for the cells. 

It is absolutely crucial to be sure that your gene is not essential since cells with full KO will be eliminated and those which retain partial or full expression will be selected. It is actually looks like your gene is indeed essential since you have 50% expression (heterozygous KO, each cell has one allele being KO). 

Hope it was clear.

Best,

Evgenii

I would also designed the sgRNAs for the functional domain (you can be sure you have a functional deltion) or in your case maybe the domain that WB antibody recoginizes

Thank you all for the help. I used the gRNAs from a library that we used for screening, although I should go and check the target. I used the 3 gRNAs separately, but I did start an experiment where I am combining the different gRNAs. I will hope that the combination of all 3 should give me better knockout. In future I will use the vector with fluorescent signal to sort the single cells instead of randomly guessing and picking from the 96 well plates. 

Sehrish, I do not get your idea about fluorescent reporter - puromycin selection sounds more than enough.

Could you answer these questions so that we can try to workout what is the problem-

Do you have massive cell death when selecting with Puro? 

At which day you see the most cell death?

Which cell line are you using and which gene you try to KO?

Which library these gRNAs come from?

I would use two sgRNAs and try to delete the functional domain and then subclone to get a homozygous deletion

I agree with Yehven, if you try to knockout essential genes then, you end up selecting only heterozygous KO cells and cells that did not have any knockout of the gene. You should find the region your clone targets before the transduction

I agree, you should not be using Western Blots for KO confirmation.  Even though you are using gRNAs from a screening library, have you confirmed that the gRNAs you are using hit all known transcripts of your gene?  Have you performed a mismatch assay to determine cleavage efficiency of the gRNAs?  Have you looked at the electropherogram data to determine heterozygosity?

MilliporeSigma can help you design your gRNAs and walk you through your CRISPR experiment.  I have attached a blank application to submit to [email protected]

Thank you all for your help. So I know for sure, the my targets are non essential genes, or they get compensated for their loss so I am not losing the cells because of the knockout. Is there a way for me to find the region my gRNA is directed to? 

Sehrish, do you lose any cells because of Puro selection? If you are using virus mix with 3 different not validated gRNAs and loosing cells because of Puro selection (meaning you have low virus titer) than what you get is not a surprise - that is what you should expect. 

This is true that you need to ensure that gRNAs cuts all transcripts but in most libraries they are already designed to target all transcripts. WB might sometimes detect inactive protein but it is really rare  since premature stop codon generated by CRISPR edition result in unstable proteins targeted for degradation.  

The most important question is - do you have cell death during puro selection? If yes, try to do single gRNA knockouts and check them with WB   OR optimize your packaging protocol to get high virus titer.

I think people could help you if you provide more details on protocol, gene, gRNAs etc.

To find targets you can BLAST your gRNAs and then check whether targeted exon present in all transcripts  using UCSC Genome Browser for example. Details on the domain functionality can be found at UniProt.

Sherish you should decide what you want in the end - do you want homozygous 100% KO or you just want stable 90-99.9% stable heterozygous KO. And then protocol should be designed accordingly. I try to advice on the current protocol you are using but it might not be the best protocol for your purposes. 

Yevhen and all,

Thanks for the help and I really appreciate this thread. I am learning so much. In answer to your queries:

1. Do I get a lot of cell death during puro selection?It depends on cell type. Some cell lines are pretty stable within first few days while others show massive cell death. I agree that I maybe selecting for the resistant clones by killing most of them. Again, in the best case scenario, I observe the loss of 30-40% of cells in 3-4 days before they start growing back.

2. Single cell knockout? I have never heard of that, can you tell me more about this?

3. Viral titre. So I am using 4ug of my lenticrispr with gRNA with 4ug of my two component packaging vector DNA(3:1 ration). I get my DNA sequenced after the preps and use the high quality DNA.

4. We avoid pooling siRNA and shRNA to eliminate off target activity. I did not pool my gRNA. I am using 3 gRNA, packed separately and I transduced my cells separately. I am gathering from this conversation that pooling them together will be a good approach. I will try it next week.

It is possible that the puro selection is in fact causing some unforeseen artifact. We use GFP expressing Cas9 plasmid from addgene, which requires 24hr post nucleofection cell sorting into single cell populations. This process is very laborious, seeking clonal cells populations however is extremely important to ensure accuracy of your genotyping/western blot analysis. The best way to ensure you are not screening a heterogenous population of cells is to clonally expand them and screen through NGS. Make sure you only have two alleles before any additional analysis (WB) due to the likelihood of generating mosaics, which may have residual expression of your protein (similar to the ranges of protein expression you mentioned in your earlier post). We prefer to use direct microinjection in parthenodes for validation of sgRNAs. This would require ordering additional oligos for annealing and processing in a T7 expression system for generating in vitro transcribed RNA for injection. You can order in vitro transcribed RNA from companies if that is an option for you. 

Hope this makes sense.

Best of luck, screening targeted cells is time consuming, stick with it.

I'd like to start with a little discussion.

In my opinion, there are two major ways of performing a CRISPR KO in cell lines :

1. Lentiviral KO performed on a cell population. This is a "quick and dirty" approach which is good for initial hypothesis checking. Drawbacks are higher off-target and heterogeneous population (not all cells in the population are KO, still in mast cases it is from 80 to 99% in my experience). However, phenotype seen in this approach is still is very strong evidence in my view.

2. "Clean" KO performed with plasmid transfection, microinjection etc. This is much more laborious approach since you typically want to subclone two times (before and after KO) and screen for homozygous KO.  

Many people advised you to proceed with "Clean" KO which is good but you need to be sure that you will observe an "important" phenotype. Still, many studies just use lentiviral KO as a final evidence. So you need to decide which approach you use. 

Let's get back to your experiments. I have done quite a few KOs and have never experienced such a low efficiency which you report. Here is a possible reason why it might happen:

Since you have 30% cell death than the multiplicity of infection (MOI) is most likely low. Therefore, you might have low concentration of gRNA/Cas9 which result in a low activity and ,thus, low cutting rate. This effect might sum up with such things as low gRNA activity and high stability of WT protein.

This what I would do if I stick with the lentiviral KO:

Check the gRNA activity on benchling. If it is ok, infect cells in MOI 10. After 2 weeks check KO with WB and if you still have the same 50% KO efficiency you could do NGS (50 000 reads on miSeq would be a cheap option) to find out the reason.

Meanwhile you could also grow subclones from your transduced  populations. In this case you will compare a single KO clone with a non-KO population. For this reason, I think that this approach is the worst unless your population was grown up from a single cell clone. 

I wonder if someone could kindly pass me a curated protocol for clonal isolation of homozygous 293T CRISPR/Case9 knock-out (dilution method preferred, not FACS). Thanks in advance!

The "cleanliness" of a CRISPR (Clustered Regularly Interspaced Short Palindromic Repeats) knockout refers to the precision and specificity with which the CRISPR/Cas system edits the genome, and the absence of off-target effects or unintended mutations. Several factors determine the cleanliness of a CRISPR knockout:

To assess the cleanliness of a CRISPR knockout, several steps are usually taken:

• Designing Specific gRNAs: Using bioinformatics tools to design gRNAs with high specificity for the target site and low likelihood of binding to similar sequences elsewhere in the genome.
• Screening for Off-Target Effects: Techniques like whole-genome sequencing, GUIDE-seq, or targeted sequencing can be used to check for off-target mutations.
• Validating On-Target Editing: Sequencing the target site to confirm that the intended edit has been made and to characterize the nature of the indel or mutation.
• Functional Assays: Beyond genetic analysis, functional assays to confirm that the gene's activity has been altered as expected can also be informative.

The cleanliness of a CRISPR knockout is a critical factor for both research and therapeutic applications, as unintended mutations can have unforeseen biological consequences. The field is continuously evolving, with advancements in CRISPR technology regularly improving the specificity and efficiency of genome editing.

l This list of protocols might help us better address the issue.