Poor transfection efficiency of PX458

[simon durand]

Hi all,

I am having trouble to transfect the PX458 vector into my cells (neuroblastoma adherent cells). I use an NaCl electroporation for these cells and I usually obtain a 90% of GFP+ cells using a reporter plasmid (pMax, promotor:CMV), but with the PX458 I have a transfection efficiency near to 0, and the few fluorescent cells are not as bright as for the pMax. I checked the sequence of my plasmid and I also digested it, it's correct. Has someone experienced something similar ? I'm not sure if this is due to the size of the plasmid (9k against 3k for the reporter) or because the Cas9 is not expressed for some reason. Could I get a better efficiency using co-transfection of MLM3636 and JDS246 ?

Thanks a lot for your help,

Simon

[Phil Abbosh]

Simon

I had a similar problem going from pmaxGFP to a GFP-expressing retroviral construct in a viral packaging cell line and found that the optimization i did for the pmaxGFP vector did not carry over when i used the larger retroviral vector (using nucleofector electroporation as it sounds like you may have).  the poorer efficiency of the larger plasmid was partially overcome by putting 5 ug of plasmid DNA into the electroporation reaction rather than 2 (i tried several DNA doses, >7 was pretty toxic).  i did not change the nucleofection program though. 

there are also several threads in this google group suggesting that the GFP in some of these multi-cassete vectors doesnt shine very brightly.  in my hands transfecting pMax as 10% of the total input DNA as a separate plasmid will result in dual transfection of about 80% of the GFP+ cells.

[Jerry Xu]

Hi Simon I have exactly the same problem

Nucleofection using pMax-GFP shows perfect results, but PX458 I don't see anything!

[RB]

I got similar results when comparing px458 with px459v2+pEGFP N1 but I think that it is the level of GFP expression from the PX458 vector that is much lower than from other vectors encoding GFP, probably due to the promoter used or the 2A autoproteolytic peptide.

Try to use px459v2 for cas9 and guide expression cotransfected with pEGFP-N1 vector or pMax.
Also try to compare px458 alone, pMax alone, and a condition with px458+pMax. If you get a lot of shiny GFP+ cells in the px458+pMax, it probably means that px458 is well transfected but that the GFP expression from the px458 is low. If you have low number of GFP positive cells in px458+pMax, it means that your px458 prep has some problem that inhibits transfection.

Hope that helps.

[Jerry Xu]

Thanks RB.

My concern is if PX458 GFP is very weak, that's very likely because the CBh promoter is NOT working well at least in my cell line. Then Cas9 is expressed at very low level, achieving very low cutting efficiency.

I don't know what's your cell line please? Also neuroblastoma?

[RB]

Hi Jerry,
In my case I transfected both NIH3T3 and mouse ES cells.
And in both case I got low GFP signal or low number of GFP positive cells with PX458 vectors. I did not try cell sorting, only fluorescence with inverted mcroscope.
However I also cotransfected px458 expressing a guide of interest and a reporter vector that express puromycin only when there is cutting of the reporter by a guide (and repair of the reporter vector). And after puromycin selection I got same cutting efficiency as with px459 vector (tested on bulk cells 3 to 4 days after stopping puromycine selection)
So from these experiments (same results in NIH3T3 cells and ES cells), I conclude that even if there is a low expression from Cbh promoter, it is enough for Cas9 to be expressed at a sufficient level to efficiently cut its target, and also it is sufficient for puromycin resistance gene to gie puromycin resistance, but maybe not enough for eGFP to be detected with our conventionnal fluorescence microscopes.

I think it is also better thats Cas9 is not too much expressed, so that you don't overload the cells with cas9.

In addition it is possible that the 2a autoproteolytic seqence is problematic in your specific cell type?

What I would suggest is that
1-co transfect px458 with a bit of pMax or pEGFP-N1 (or pEGFP-C1) and check 24h later. If you get a lot of cells expressing GFP, then it means normally that the transfection is ok and that it is not the reason for low gfp with px458. (If px458 is inhibiting transfection, then pMax should also not be transfected, and then you will get low number of gfp positibe cells, at least using lipofecyamine, I don't know how co-transfection works with nucleofection or electroporation).

2-you co transfect px458 with puromycin expressing vector and select with puromycin 2 days then let the cells grow and check for cutting in the bulk population. If with at least 3 different guides you don't get any cutting then there is probably a problem with cas9 expression (that could explain low or poor gfp).

3-you can transfect px458 and you can try to stain your cells with anti-Flag and/or anti-GFP to check if cas9 and gfp are expressed by your px458 pasmid.

4-try also to check your cells transfected with px458 under the fluorescence microscope using PBS in your well instead of the culture media (sometimes it allows to show low GFP expressing cells)

5-I assume that you want to use Px458 in order to FACS-sort the cell, and thus I think that the only way to get decent GFP signal for sorting is to cotransfect px458 with anothr vector encoding GFP (pMax or pEGFP).

I hope that these small things will help.

Keep us updated with your progress :)

[Patrick Weber]

I also use pX458 for my experiments and recently transfected mouse myoblasts (C2C12) as well as mES cells using PEI.

Indeed I could barely see (by epi-fluorescence) any GFP expression in either cell type. Only in case of C2C12 I could find some cells with relatively low expression.. wondering why this is the case I stumbled over this post.

FACS analysis, however, also only showed rather low transfection efficiencies (~2%). Co-transfecting pMax in lower amounts might help sorting generally more GFP+ cells.. also the puro-thing sounds like a good idea.. may also try it ;)

Best

[Jerry Xu]

Hi RB
Thanks for such detailed analysis.

I'm actually using PX459, but PX459 cut well in HEK cells but interestingly  NOT in my neuronal cell lines.
I nucleofected PX459 into neurons, wait either 24 hrs or 48 hrs, then selected for two days, then all cells died, including both CRISPR group and control.
In order for troubleshooting, I transfected PX458 (nucleofection) because PX458 only differ at GFP part from PX459. When I found GFP is barely expressed in my neuronal lines, this is quite consistent with puromycin failure; very likely the expression of Cas9+puro or Cas9+GFP is simply too low in my system.

I've tried as you suggested mix PX458 and pmax. transfected using lipo-based reagents into neuronal line (efficiency may not be that high as compared to nucleofection); PX458+PMAX show very shiny GFP, indicating PX458 plasmid prep doesn't cause transfection problem.

All the above makes me think it's promoter problem.

Btw what approach you use for bulk DNA? I simply sequence them, expecting noise caused by indel if there's genome editing, supposing NHEJ efficiency is quite high

[Jerry Xu]

btw

tested on bulk cells 3 to 4 days after stopping puromycine selection

I'm confused by this.

I typically wait 24-48 hours after transfection, then add puromycin for 3-4 days, harvest cell and genotyping.

You'll continue wait 3-4 days after 3-4 days selection? That's add up to 8 days

Also how long it takes for Cas9 to be effectively cutting?

On Tuesday, December 8, 2015 at 3:44:16 PM UTC-5, RB wrote:

[RB]

Hi Jerry,

For analysis of bulk cell, I run a normal PCR that gives an average 200bp product, centered on the cutting site. I run this PCR on a 2.5% agarose gel. Normally if you have indels you should get heteroduplexes formation which will give you about 2 bands on the gel (1 band similar to WT and another one 5-10b higher due to heterodulexes). You have to run the gel for quite long.
However it is hard to quantify efficiency with this method.
Your method of sequencing is also good I think. I have not tried yet.

For my transfection protocol, 24h after transfection, I start puromycin selection for 48h, then I stop selection and keep culturing for 3 to 4 days (just to amplify and get a lot of DNA).

I never did any kinetic to know when cas9 starts cutting but I believe that 24h after transfection, it should have already cut (and this is corroborated by my assay with the puromycin reporter that is active already at 24h).

Did you check that the target sequence in your neuroblastoma cell line is not mutated or does not contain SNPs in wild type cells?

What you could try is to modify the Cbh promoter of px459 by a CMV or sv40 promoter (by cloning or recombineering)

[Rafael Maciel ioris]

I am very disappointed with plasmids px458 and px459.

pX459 does not give any puromycin resistance! So, how I am supposed to select??

We used the second generation. The one that was supposed to be correct. We used two different ways to transfect, polyplus and lipofectamine. We used two different cancer cell line, A549 and MDA-MB-231. We used several different time points to start the selection after transfection. We used several different concentration of puromycin. And nothing worked! Addgene do not take any responsibility. At the end, this group is the unique way to find help.

pX458 has very low transfection efficiency! 

We used two different ways to transfect, polyplus and lipofectamine. We used two different cancer cell line, A549 and MDA-MB-231. We used several different time points to start the selection after transfection (in this case we use FACS). We obtained 0.02% efficiency with MDA-231 and 0.1% with A549.

As far as I could read in this group, many many others are having the same trouble using different ways to transfect. So, I am wondering, is our protocols all wrong?

[Tim Vierbuchen]

Hello,

1. The plasmids are totally fine - I'm sure that the cell line is your problem!
2. Do you know if the Cbh promotor is active in your cells? Addgene has also CRISPR plasmids with a CMV promotor. Might be better for your cells.
3. Please consider that your cell line might be hard-to-transfect. Our group is working with another bronchial epithelial cell line (Calu-3). We only get sufficient transfection rates using electroporation. Lipofection works really bad for them...I guess the mucus layer inhibts the transfection. We are using the nucleofector device from Lonza.

[Jonathan Geisinger]

Question:  What are you sorting on?  pX459 doesn't have a FACS-able marker on it.

[crispr.m...@gmail.com]

The plasmids work, and in many cell types in my experience. As others have pointed out it's probably a problem related to your cells. Do they routinely transfect with other plasmids? Have you checked for mycoplasma recently?

On a seperate note the plasmids are provided through addgene, a not-for-profit organisation, for a relatively inexpensive fee in the spirit of collaboration. Perhaps if you're not happy with the services and reagents provided you could make your own?

[Phil Abbosh]

Hi Rafael
i think MDA231 are hard to transfect. you might at least check if using a GFP vector.
Mycoplasma may also be a problem. strongly suggest checking that if you dont already.
the protocols arent wrong though. many have had trouble, but many more have had success.  Addgene is a wonderful service for all of us but its up to the individual scientists to troubleshoot.

if your cells are hard to transfect or you want to try another approach, try lentivirus.  I made a version of lenticrisprV2 that has replaces the codon-optimazed SpCas9 with eSpCas9(1.1).  it is Addgene #78852 if you are interested. We have used it in about a dozen knockouts with success every time, single cell knockouts in 1 month. Several other users have given me feedback that it works just fine. you can use the same oligos you designed and purchased to clone into X459 - #78852 uses the same overhangs. 

phil