1. Always around the cut site.  Your left arm of homology should end 3 nucleotides before your Pam motif. Your right arm of homology should start 3 nucleotides before your Pam motif. Use 60-70 bp for each and order your final construct as a 4nmole IDT ultramer.

2. Yes, your cut can lie a few (5 nt) outside the arms. The closer the better to avoid NHEJ.

(1) We've only done two-cut experiments so far (GFP insertions -- one cut on either side of the desired insertion), but it sounds like you'd want the homology arms flanking the cut site, not the point mutation. The protocols I've looked at have all recommended starting the homology arms *immediately* at the cut site. But I don't know how rigorously this has been examined.

*Our work is in Drosophila, so the particulars might be a bit different in your case. At least in fruit flies, we're using 1kb homology arms. Assuming that you're using a similar length, the 50bp between the point mutation and cut site should be fairly irrelevant.

(2) If you run a mutagenesis PCR (silent mutation) to change the final G in the PAM sequence, you won't have to worry about re-cutting. That said, our lab has seen success even without mutagenizing the CRISPR target. But theoretically, yes, it could be a problem, so mutating the PAM seems like the safe bet. (The Q5-SDM kit has worked great!)

Daniel,

Why do you do 2 cuts when knockins only require 1 cut? (I'm just curious)

We do a pretty large insertion (700bp GFP, plus 1250bp for a floxed selectable eye marker). Doing two cuts allows us more flexibility.

This way we can put the GFP right at the 5' or 3' end of the gene, and keep the selectable marker further away from the endogenous gene, to minimize disruption of the UTR. (Generally around 1kb up/downstream from the stop or stop codon.)

We're working from a modified version of the protocols I've linked below.

http://flycrispr.molbio.wisc.edu/protocols

The closer the center of the point mutation in your HDR donor to the crispr cutting site, the higher knock-in efficiency. However, 50 bp away from the cutting site is not a big deal to me if you use long homology arm (500-1000 bp). But if you want to use ssOligo as a donor, it could be a big problem. 

According to Dr. Zhang:

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 (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 further information please:

 http://www.addgene.org/crispr/zhang/FAQ/

Thank you guys for your feedback - I appreciate the insight and help.

Dear Patrick

Hello,

dissociation of Cas9 from double-stranded DNA (dsDNA) substrates is slow (lifetime ~6 h) but that, before complete dissociation, Cas9 asymmetrically releases the 3'-end of the cleaved DNA strand that is not complementary to the sgRNA (nontarget strand). By rationally designing single-stranded DNA (ssDNA) donors of the optimal length complementary to the strand that is released first, scientists increase the rate of HDR in human cells when using Cas9 or nickase variants to up to 60%. 

Please see the following paper:

http://www.nature.com/nbt/journal/v34/n3/full/nbt.3481.html

Best,

In CRISPR-Cas9 genome editing, particularly when creating point mutations, the "center" of your Homology-Directed Repair (HDR) construct is an important consideration for efficient and precise editing. The HDR construct typically contains a homology arm sequence that is identical to the target DNA sequence, except for the desired point mutation.

For a CRISPR-Cas9 HDR construct aimed at introducing a point mutation, the "center" is ideally positioned around the point mutation itself, rather than the Cas9 cut site. Here's why:

In summary, when designing an HDR construct for a point mutation using CRISPR-Cas9, the point mutation should be near the center of your construct, flanked by homology arms, and ideally, the Cas9 cut site should be close to but not overlapping with the mutation site. This positioning optimizes the likelihood of successful and precise genome editing.l This protocol list might provide further insights to address this issue.