Correction to my post "Protocol for cloning protospacer adapters...PX330-family #CRISPR plasmids".

There was an error - now corrected - in my previous protocol post  "Protocol for cloning protospacer adapters into Zhang lab PX330-family #CRISPR plasmids.":  I omitted by mistake the instruction to add the ligase enzyme to the ligation reaction. (line II - 2 - e in the protocol).  Duh!    I apologize profusely.

Hard numbers re: eGFP and CRE knock-ins with #CRISPR in rat zygotes!

This one has important implications for CRISPR-mediated insertion of fragments in the several-kilobase size range, such and GFP, CRE, etc. , into rats - and likely, mice - via pronuclear injection.  

Generation of eGFP and Cre knockin rats by CRISPR/Cas9.   FEBS Journal.  Accepted manuscript online: 17 JUL 2014.   

Yuanwu Ma¹, 

Jing Ma¹, 

Xu Zhang¹, 

Wei Chen¹, 

Lei Yu¹, 

Yingdong Lu¹, 

Lin Bai¹,

Bin Shen², 

Xingxu Huang^2,* and

Lianfeng Zhang¹

A brief summary from a quick dig I made into this paper:  They did 3 different CRISPR targeted kncok-ins, that is, using HDR to insert cassettes into genes of interest.  The cassettes were GFP and CRE into 1 and 2 different genes respectively.  This was all done by pronuclear injection into rat zygotes, of RNAs for Cas9 and guide RNA plus circular, double-stranded DNA plasmids containing the cassettes of interest.    So this is technically, essentially the same process one would use for mice.   Their numbers were quite impressive:   between 23% to 54% of live pups carried the targeted insertion.

A key ratio to scrutinize is the ratio of targeted insertions to the total number of insertions and other mutations, e.g. indels.   The reason is that the total number reflects the number of pups in which CRISPR clearly had activity.   Therefore the ratio reflects the proportion of the time that the HDR process occurred successfully as a subset of all embryos that had some sort of CRISPR-mediated cleavage event.  Since almost all of their live pups had evidence of the HDR insertions or indels anyway, the ratio of targeted/all events is still about 25-50%.   

How big were the homology arms used?   Between ~1.5 to 2.1 kb in all cases (3 constructs x 2 arms each = 6 arms total).    Concentrations used were:   25 ng/µl Cas9 mRNA, 10 ng/µl, and 4 ng/µl of the donor plasmid in circular form.   

New paper: Mosaicism and complexity in #CRISPR founder mice.

Yen et al. recently published a paper in Developmental Biology (Yen at al, 2014) with some important observations about founder mice generated from injecting CRISPR tools into mouse zygotes.

They targeted Tyrosinase (Tyr), which causes albinism in the homozygous-null state and thus an easy readout of CRISPR function to mutate this gene.   Like others, they observed high rates of success with many fully albino mice being generated, indicating no surviving wild-type alleles.   However, they note many genetically mosaic animals were born too - these had patches of white fur among the pigmented fur, indicating clusters of homozygous-mutant cells, among other patches of cells that clearly still had wild type Tyr function.   This clearly suggests a high rate of CRISPR mutation that did not occur till after the first embryonic cell division.  They also did sequencing on the live born founders to determine the new mutations in the target gene.  This revealed that the mice could clearly contain more than two types of detectable new mutant alleles.  In fact, they found 57 mutant alleles in 23 total mice!    So, determining the exact mutagenic outcome in a founder animal is complex.   Furthermore, founder mice could potentially transmit more than two types of mutant alleles to their progeny, presuming their germline will also be mosaic.   Of course, F1 progeny must be carefully screened and sequence-validated to figure this out.

Current thoughts on targeting reporter cassettes in mice using #CRISPR.

 As an advocate of the PX330 plasmid for implementing CRISPR in mouse zygotes, I am trying to figure out optimal conditions for enabling homology-dependent repair (HDR) with this system.  Genome editing requires HDR, and a donor DNA molecule must be supplied along with the CRISPR/Cas tools such as PX330.  

For introducing "smallish" edits or insertions - like, under 70 bases or so - one can order a single-stranded super-long oligo such as IDT's ultramers.  Larger insertions are going to require double stranded DNA fragments.  Most molecular biologist are well used to purifying these, once they are designed.   Here's a few considerations:

1.  If the CRISPR target is also present in the homology arms of a double-stranded donor, it's probably gonna cut the donor before it gets a chance to donate anything.

2.  What should the optimal ratio of donor/CRISPR/Cas9 be?    I am assuming roughly a 1:1 mass ratio.   My only guidance so far is from the Yang et al 2013 Cell paper.  They were coinjecting donor DNAs with RNAs for guide RNA & Cas9.   Here I will focus on their data for fluorescent reporter fragment insertions, which were injected as circular dsDNA plasmids.   Although this group mainly does cytoplasmic injections, they tested pronuclear injections too.

For pronuclear injections of these reagents. Yang et al  reported 9% and 18% targeting rates, using two different donors and targets.    The concentrations were:  10 ng/µl donor plasmid, 2.5 ng/µl guide RNA, 5 ng/µl Cas9 mRNA.    Thus the total nucleic acid burden of the pronuclear injection material was ~17 ng/µl.   Those with actual experience with mouse embryo injections will note this is rather high;  most DNA transgenes are injected at 1-5 ng/µl, no more.  Why no more?   It can cause toxicity, but even more often and more frustrating, it leads to frequent needle clogging at the injection microscope.   Injected material at these concentrations will need very careful preparation beforehand.  I suggest that a 0.22 micron spin filter be used before each injection over a total 5 ng/µl limit.  Anyway, a ratio of ~1:1 or maybe ~3:2 mass ratio of donor DNA vs. CRISPR reagents seems reasonable.    For PX330 I am starting off by suggesting 4 ng/µl PX330 + 6 ng/µl donor plasmid - although I have no data yet showing whether this is optimal.

Low rate of #CRISPR off-target mutations in human iPS cells, reported in 2 new papers.

Veres A, Gosis BS, Ding Q, Collins R, Ragavendran A, Brand H, Erdin S, Talkowski ME, Musunuru K.   Low Incidence of Off-Target Mutations in IndividualCRISPR-Cas9 and TALEN Targeted Human Stem Cell Clones Detected by Whole-GenomeSequencing.  Cell Stem Cell. 2014 Jul 3;15(1):27-30.

Smith C, Gore A, Yan W, Abalde-Atristain L, Li Z, He C, Wang Y, Brodsky RA, Zhang K, Cheng L, Ye Z.   Whole-Genome Sequencing Analysis Reveals HighSpecificity of CRISPR/Cas9 and TALEN-Based Genome Editing in Human iPSCs.  Cell Stem Cell. 2014 Jul 3;15(1):12-3.

These papers are very important for using WGS to thoroughly catalog all variants in iPS cells post-CRISPR (and TALENs).  Good news:  Very very low rate of off-target (OT) CRISPR mutations, in contrast to some previous reports of high OT rates in transfected cells.   The authors suggest that the discrepancy may exist because the other studies used different, more commonly-used, "workhorse", non-stem, transformed/quick replicating cell lines.  It's possible that there are some technical differences in transfections and/or specific CRISPR reagents that contribute to these differences, but it is nice to see two different groups in agreement on the iPS situation.  Also, this is reminiscent of the observation by several groups that CRISPR OT effects in mice (generated by zygote injection of CRISPR reagents) are also very minimal.

So this is the not-so-good news, not for CRISPR per se, but for clonal propagation of iPS cells in general:  Both groups discovered that iPS clones accumulated numerous non-CRISPR-related new mutations.  That is, the act of isolating and passaging clonal cell lines itself led to accumulation of 50-100 new single-nucleotide variants not seen in the parental cell line.  This is genome-wide, so only a few of these are likely to be within exons, but still.     The bottom line is that iPS subclones are not, strictly speaking, genetically identical to the parent cell or each other.  Whether this is going to be a major problem going forward in the iPS field remains to be seen.

Protocol for cloning protospacer adapters into Zhang lab PX330-family #CRISPR plasmids.

This is also publicly accessible at the Vanderbilt labnodes website.

Cloning protospacer adapters into PX330/PX335/PX458/PX459 (Zhang lab, MIT) via BbsI site.

I.               Anneal top and bottom oligos of adapter

1.     Resuspend oligos in water to 1 µg/µl concentration.

2.     Combine the following in a 200 µl PCR tube:

a.     5 µl “top” oligo

b.     5 µl “bottom” oligo

c.     10 µl of 10x annealing buffer

d.     80 µl water

(10X Annealing Buffer: 1 M NaCl / 10 mM EDTA pH 8.0 / 100 mM Tris pH 7.5)

3.     Mix well.  Run the mix in a thermal cycler with an annealing program, such as:  Step (1) 94˚ for 3 minutes; Step (2) cool from 94 ˚ to 25˚ slowly, such as over a 30 minute period.  Alternatively to a thermal cyler:  Heat a beaker of water to boiling. Float the tube in the boiling water bath for 5’. Remove the beaker from the heat and let it cool off naturally on a benchtop, till it is room temperature.

4.     Transfer the annealed adapter to a 1.5 ml tube.  Add 900 µl water.  The final concentration of annealed adapter is now ~ 10 ng/µl.

II.              Ligate to BbsI-cut vector.

1.     Before you start:  You will need to have the vector DNA previously cut with BbsI and gel-purifed, and in 10 mM Tris or Lo-TE at a concentration of at least 10 ng/µl.

2.     I use the NEB Quick Ligation kit, and I reduce the volumes by half to save reagents.  It works great.  Combine the following in this order:

a.     2.5 µl of 10 ng/µl BbsI-cut vector

b.     1 µl of 10 ng/µl annealed adapter

c.     1.5 µl of water

d.     5 µl of 2x Quick Ligase buffer

e.   0.5 µl of Quick Ligase enzyme

3.     Mix briefly, and incubate at room temp. for 5’.  Use immediately for transformation.

III.            Transform DH5a cells.

1.     You will need a 42˚ water bath and LB+AMP plates.  You will also need to get a tube of competent DH5a cells from the 9^th floor core in Light Hall.  Thaw the cells on ice.

2.     Transfer 100 µl cells to a 1.5 ml tube.

3.     Add 5 µl of the ligation reaction. Mix well (do not vortex).

4.     Incubate 45’ on ice.

5.     Heat-shock the cells at 42˚ for 2’.

6.     Transfer cells back to ice.

7.     Add 900 µl of LB media to the cells. Transfer the cells to a 15 ml tube.

8.     Recover the cells by incubating in a 37˚ shaker incubator for 30’ at 250 rpm.

9.     Plate out 100 µl of the cells on an LB+AMP plate. Incubate at 37˚ overnight.  You should get at least a few dozen colonies (e.g. 10-100 is typical). Most of them will contain the correctly cloned product.   When I have done negative controls with no adapter, I get zero or just a couple of colonies.

IV.           Screen colonies for correct insertion of adapter.

1.     Inoculate 2 colonies separately into 15 ml tubes with 2 mls of LB+100 µg/ml AMP.  Shake overnight at 37˚.

2.     Perform Qiagen minipreps on 1.5 ml of each plasmid culture.  Save the remaining ~0.5 ml for making a glycerol stock to store at -80˚. (see below).  The culture sample can be stored at 4˚ for a few days before you make the glycerol stock.

3.     Spec the DNA.  Usually the concentration is about 40-100 ng/ul.    Prepare a sample for direct Sanger sequencing using this primer:

U6F1:              TACGATACAAGGCTGTTAGAGAG

This sequencing will read-through the region of the BbsI site and verify that the adapter has inserted properly.

V.             Make glycerol stocks of the correct clones.   To the remaining ~0.5 ml of the miniprep culture in step IV above, add 0.5 ml of sterile 30% glycerol.   (Glycerol solutions should be sterilized by filtration, never autoclaving.) Mix and store the glycerol stock at -80˚.

Doug Mortlock 7/2/14