Poor gibson efficiency in lenti CRISPR V2

[Herschel Dhekne]

Hello Experts,

I performed a Gibson reaction with my custom library and lenti guide puro as well as Lenti CRISPR V2. I got 65 fold more colonies in lenti guide puro than control reaction. But for lenti CRISPR V2, I got only 7 fold more colonies than control for some reason. Joung et al Nature Protocol 2017 paper recommends 20 fold more colonies.

Should I repeat the whole thing or should I sequence the maxi prep and see how the library looks?

Thanks everyone for your inputs.

Herschel

--

Greetings,

Herschel Dhekne

[Agnieszka Krawczyk]

Hi Herschel,

In our hands Gibson also gave very poor efficiency for cloning custom library into lentiCRISPRv2. Instead, we tried circular polymerase extension cloning (CPEC) method and it worked great, ca. half million colonies from one electroporation. I don't remember now how about the negative control exactly but certainly much less. And the whole library was perfect, no missing guides, low skew ratio. Perhaps you may give CPEC a try as well.

Best,

Agnieszka

[Anthi Demetriadou]

Hello Herschel and Agnieszka,

I have several issues with my custom library as well and I believe that the reason is the low efficiency of my Gibson reactions (I used lentiGuide-puro as backbone).

I have explained in detail what I have done here https://groups.google.com/forum/#!searchin/crispr/julia$20joung|sort:date/crispr/qBrCz38vW18/O-fRVUR-AwAJ

Herschel, did you change anything in the Joung et al Nature Protocol 2017 or did you follow exactly the protocol? How big is your custom library? Which electrocompetent cells did you use?

Agnieszka, can you share with us the protocol with CPEC method? Maybe I can give it a try...

I will appreciate your help!

Thank you,

Anthi

[Agnieszka Krawczyk]

Hi Anthi,

I just read your story in the other thread and it very much resembles issues we had with Gibson cloning. Please find attached our adapted protocol for CPEC, based on this paper PMID: 24395360. We used it for cloning a ~45,000 sgRNA library into lentiCRISPRv2 and ~8,000 sgRNA library into lentiGuide puro, both looked perfect after NGS. You can use for it exactly the same oligos, amplified according to the Joung et al, Nat Prot 2017. Hope it helps, let me know if you have questions. Good luck with your cloning!

Best,

Agnieszka

[Julia Joung]

Hi Herschel,

It's odd that you got lower fold change in colonies for lentiCRISPRv2 than lentiGuide Puro, because I would have expected the reverse.

That being said, it's probably okay to proceed with maxiprep and see how the library looks - library might be totally fine! You may want to do a test digestion using Esp3I on the maxipreps for both libraries just to make sure the proportion of background vector is low before sequencing.

Best,

Julia

--
You received this message because you are subscribed to the Google Groups "Genome Engineering using CRISPR/Cas Systems" group.
To unsubscribe from this group and stop receiving emails from it, send an email to crispr+un...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/crispr/b2f5a433-1605-4d07-a15b-531743c9edb6%40googlegroups.com.

[Anthi Demetriadou]

Dear Agnieszka,

thank you very much for providing your protocol. I will give it a try and I will let you know!

Thank you again!

Best,

Anthi

[Anthi Demetriadou]

Hi Agnieszka,

I was wondering if you use a control CPEC reaction for electroporation, for example a CPEC reaction without insert to compare the electroporation effieciency between the CPEC reaction and the control CPEC reaction. If the answer is yes, how many colonies it is reasonable to have in the control electroporation to consider your CPEC reaction successful?

Thank you,

Anthi

On Tuesday, March 10, 2020 at 8:05:07 PM UTC+2, Agnieszka Krawczyk wrote:

[Agnieszka Krawczyk]

Hi Anthi,

We cloned two libraries using CPEC and each time in the negative control CPEC reaction we had ca. 10% of the colonies observed in the library CPEC reaction. We checked by colony PCR and sequencing 25 colonies from each. It's a small sample but in negative control two colonies were uncut vector and the rest didn't give products. In the library, one was uncut vector, another didn't give product, and the rest was sgRNAs from our libraries. NGS showed very good results for both libraries. How did your CPEC go?

Best,

Agnieszka

[Anthi Demetriadou]

Hello Agnieszka,

I will try the CPEC method next week.

What I have in mind is to initially perform CPEC with the vector (lentiGuide-puro cut with Esp3I) and PCR-amplified oligo library that I used before to perform the Gibson reactions. I will follow the second page of your protocol and I will perform 3 library CPEC reactions (lentiGuide-puro, Esp3I + PCR-amplified oligo library) and 3 control CPEC reactions (lentiGuide-puro, Esp3I only). What do you think?

Basically, I want to check first the electroporation efficiency of the CPEC reactions on the NEB electrocompetent cells that I have in order to decide how many CPEC reactions and electroporations are required for my library.

Thank you for the advice regarding the number of colonies that I have to expect from the control CPEC reaction and for your help in general!

I will let you if I success with this!

Best,

Anthi

[Agnieszka Krawczyk]

Hi Anthi,

3 reactions for a trial seems ok. The least we did was 5 reactions. Just remember to resuspend them after isopropanol precipitation in no more than 1-1.5 ul * number of reactions, so that you have high enough concentration to be able to use 300 ng for electroporation (no more than 2ul per electroporation). Good luck!

Best,

Agnieszka

[Anthi Demetriadou]

Hi Agnieszka,

this is to let you know that my CPEC electroporations went very well.

I performed 3 electroporations using 200ng per electroporation of the CPEC reactions and I got a very good efficiency according to the size of my library.

I will perform colony PCR followed by sequencing, as you suggested, to check a number of colonies from both the CPEC and control reactions before I proceed to midipreps and NGS.

Thank you again for sharing your protocol and your experience!

It was really valuable for me!

Best,

Anthi

[Agnieszka Krawczyk]

Hi Anthi,

I'm happy to hear that CPEC worked well for you. Hope your library is fine and good luck with your further experiments!

Best,

Agnieszka

[Minoo karimi]

Hi Agnieszka,

I tried your CPEC protocol  to make a custom sgRNA library with 100,000 oligos. Unfortunately the reaction did not work. I am using NEBNext Ultra II Q5® Master Mix . 

I used 100 ng of vector (LentiCrispr V2 plasmid), 10 ng of my oligos (prepared based on the Nature protocol). I prepared 6 reactions, precipitated them by isopropanol. I electroplated 300 ng/ 25 ul of commercial competent cells.  incubated 1 hour at 37 C. Polled the reactions and plated with several dilution. 1X, 10X, 100X and 1000x but did not see any colonies :( 

How many colonies do you usually get per electroporation? I do appreciate any suggestions. 

Thanks,

Minoo

[Minoo karimi]

Hi Anthi,

I was wondering if you got your NGS results. Did CPEC worked? Could you please share your protocol with me? I am having trouble with Gibson reaction as well as CPEC.

Thanks,

Minoo

[Anthi Demetriadou]

Hi Minoo,

Yes I did the NGS and the results were very good. You can see my NGS statistics below:

Number of perfect guide matches: 548045

Number of nonperfect guide matches: 22696

Number of reads where key was not found: 12868

Number of reads processed: 583609

Percentage of guides that matched perfectly: 96.0

Percentage of undetected guides: 0.0

Skew ratio of top 10% to bottom 10%: 1.43

My library has 1250 guides and I used lentiGuide-puro for cloning.

I attached the protocol but you have to change some things based on your library size and vector you are using. I followed basically, Agnieszka's protocol.

Good luck, 

Anthi

[Minoo karimi]

Thanks a lot Anthi! I will try your protocol. 

Minoo

[Jordi Martinez-Quintanilla]

Dear Anthi, 

I understand that you used the Python script from Julia's nature protocols paper. This script gives you the information of perfect and nonperfect matches and skew ratio. Shall we take into consideration guides that are over or down represented in the maxiprep in order to interpret the results from the future screenings? How do we know which specific guides are?

Thanks, 

JORDI

--

You received this message because you are subscribed to the Google Groups "Genome Engineering using CRISPR/Cas Systems" group.
To unsubscribe from this group and stop receiving emails from it, send an email to crispr+un...@googlegroups.com.

To view this discussion on the web visit https://groups.google.com/d/msgid/crispr/1bc85ff1-3679-42b6-8a6b-adf24fd8b10ao%40googlegroups.com.

[Agnieszka Dzikiewicz-Krawczyk]

Hi Minoo,

Sorry to hear that you have troubles cloning your library. We used CPEC for two libraries :~46k sgRNA and 8k sgRNA. Both times it was fine. A few questions:

- was amplification of your oligos fine, did you get expected band?

- was everything fine with electroporation? Was there no arcing (hard to describe but it's a specific sound when there is arc and then electroporation doesn't work)? This may happen e.g. if you have high amount of salt in your ligation mixture.

- did you use some positive control for electroporation?

- a trivial thing but sometimes strange things happen - were your plates with ampicilin for sure and not other antibiotic?

I hope you can find solution and get your library cloned.

Agnieszka

[Minoo karimi]

Hi Agnieszka,

Thanks for the reply. To answer your questions:

1- I checked the sgRNA band at 2% agarose  gel and confirmed 140 bp. 

2- I used isopropanol precipitation to get ride of salts and then did electroporation. Time constant was 5.3. It is difficult for me to say if there was a strange sound or not since I do not use electroporation frequently. But I had a positive control which worked perfectly.

3-Yes, I am sure about antibiotic since my positive control worked well. 

The only thing I am suspecting is that I used 10ul water to re-suspend the DNA pellet and used all 10 ul in one eletroporation reaction. Anthi send me a protocol which suggested 1ul water for 1 CPEC reaction. I also realized that I am not supposed to use more than 2.5 ul of DNA in electroporation reaction. 

I will try it tomorrow again and give you an update. 

I also have a question for NGS, what is the minimum read required per sgRNA? In the Nature paper it is suggested to use more than 100 reads per sgRNA, I was wondering if 50 or 75 is OK or not? 

Thanks a lot for the help!

Minoo

[Anthi Demetriadou]

Hi Jordi,

the bioinformatician who helped me with the python script, gave me a csv file that shows the number of reads for each one of my guides.

At Box 6 of the Nat Prot paper, the last question is about the quality of the cloned plasmid library. For a library to be considered as uniform, the difference between the 90th and the 10th percentile should be <10-fold.

However, I am not sure if a not uniform representation of some guides (for example, 700 reads vs 200 reads) will affect your screening afterwards. 

I suppose that if you have enough representation of your guides in your plasmid library (>100 reads per guide) and if a proper representation of your library is maintained during screening, I think that it should be fine.

Maybe Julia Joung can give a clarification on this.

Best,

Anthi

[Agnieszka Dzikiewicz-Krawczyk]

Hi Minoo,

10 ul for electroporation? That may be the issue, we never used more than 2 ul. It's better to resuspend in smaller volume and have higher concentration, so that you can add less to the bacteria. Hope it works with the smaller volume.

As to sequencing, I think 50 reads per sgRNA is too little. I would not go below 100. We sequenced our libraries at 500-1000x, similar as the coverage we used in the screen (it was around the same cost as doing MiSeq in our case). But I think it's more important that you get good coverage while cloning your library. 100x at least. We managed 170x for our 46,000 sgRNA library, and it was perfect, no construct missing. So if you don't have a good coverage from cloning, then sequencing at high depth will not change anything.

Best,

Agnieszka

[Julia Joung]

Hi Jordi,

Following up on Anthi's comment, the library quality metrics we provide are references for determining if your library is good enough to proceed.

As for whether you take into account the over/under-representation of guides during screening, we always have a control in the screen that accounts for this. For instance, a vehicle-treated condition for a drug-treatment screen, cells harvested at day 2 for a 14-day depletion screen, etc. Then we take the relative sgRNA count between the test and control conditions for downstream analysis.

Best,

Julia

--

You received this message because you are subscribed to the Google Groups "Genome Engineering using CRISPR/Cas Systems" group.
To unsubscribe from this group and stop receiving emails from it, send an email to crispr+un...@googlegroups.com.

To view this discussion on the web visit https://groups.google.com/d/msgid/crispr/c48fbb8a-bb75-4c16-ae9d-11db54d46ef8o%40googlegroups.com.

[Minoo karimi]

Thanks very much  Agnieszka. I am doing transformation today following your suggestions. 

Regards,

Minoo

 

[Canhui Yi]

Dear  Anthi,

May I please ask you a question.  How did you get the NGS statistics?

Did the NGS company help you to analyse the data or you analyse the data by yourself with some software?

Since I amplify my pooled sgRNAs obtained from addgene followed by NGS to determine the distribution of amplification.

I think I will get raw data of NGS. So, I hope to know how to get the statistics.

Bests

Canhui 

[Jordi Martínez-Quintanilla]

Hello Canhui,

I used this web-site tool at it works very well to analyse the distribution of my guides in the maxiprep as well as the screening data:

http://pinapl-py.ucsd.edu/

Yours,

JORDI

Sent from my iPhone

On 28 Oct 2020, at 05:08, Canhui Yi <ch...@connect.hku.hk> wrote:



--

You received this message because you are subscribed to the Google Groups "Genome Engineering using CRISPR/Cas Systems" group.
To unsubscribe from this group and stop receiving emails from it, send an email to crispr+un...@googlegroups.com.

To view this discussion on the web visit https://groups.google.com/d/msgid/crispr/34ff441f-c5af-49e8-9524-26d7a9bfe005n%40googlegroups.com.

[Anthi Demetriadou]

Hello Canhui,

I analyzed my NGS data using the count_spacer.py (Python script) according to "Genome-scale CRISPR-Cas9 knockout and transcriptional activation screening" paper of Joung et al. Nat Prot 2017.

A bioinformatician can help you.

You can follow the Steps 38 & 39 of the paper.

Good luck,

Anthi

[Anthi Demetriadou]

Hello Julia,

Thank you very much for providing us with your expertise on CRISPR screens.

I have a number of questions as well.

In the case of a FACS screen and a cell line that it will undergo differentiation prior to sorting, 

1. What type of control should account for the over/under-representation of the guides?

2. Do you think that the controls below are reasonable? 

• Input (plasmid DNA library) – All gRNAs present in the library
• Undifferentiated – gRNAs that target genes essential for the expansion phase of the cells will be depleted (this population is the same to the remaining cells come from the puro selection week)
• Differentiated but not sorted - gRNAs that target genes essential for the differentiation of the cells will be depleted (Unsorted control)

3. Do you suggest any other type of control samples?

4. For the adjustment of the sorter's settings, I will use 2 types of controls; 

    Cells transduced with a gene known to be a positive control for the phenotype I'm looking for and cells transduced with a non-targeting control. Is that correct?

Any other suggestion or comment related to a FACS screen is more than welcome.

Thanks,

Anthi

[Julia Joung]

Hi Anthi,

The controls you listed all sound reasonable, but might be a bit of an overkill to analyze all of the controls. I would recommend sequencing the plasmid DNA library, differentiated but not sorted, and a sorted control. If you are sorting for the highest 10% fluorescent cells, for example, the sorted control is typically the lowest 10% fluorescent cells. During analysis, the guide enrichment is calculated as the NGS count in the highest 10% divided by the lowest 10%. For the undifferentiated cell condition, you can freeze cell pellets just in case and only process if you think it will be helpful.

Your controls for adjusting the sorter's settings sound good.

Best,

Julia

--
You received this message because you are subscribed to the Google Groups "Genome Engineering using CRISPR/Cas Systems" group.
To unsubscribe from this group and stop receiving emails from it, send an email to crispr+un...@googlegroups.com.

To view this discussion on the web visit https://groups.google.com/d/msgid/crispr/9989eaa4-8352-4479-9b9f-a4fda9751425o%40googlegroups.com.

[Anthi Demetriadou]

Thank you very much Julia!

Best,

Anthi

[Canhui Yi]

Dear JORDI,

As I use web-site tool you mention to analyse the distribution of my guides in the maxiprep, the progress bar always stop at 99.5%-99.9%, never 100%. 

I try many times. I hope to know how to address this problem.

Bests

Canhui 

 

[Minoo karimi]

Dear  Agnieszka  and  Anthi,

I would like to thank you for your great comments on the guide RNA library preparation. Finally, I managed to get 3M colonies per electroporation following your CPEC protocols. I tried the Gibson assembly following the Nature protocol and got 5.8M colonies per electroporation.  Considering the cost of the Gibson assembly, I believe CPEC is a great alternative. 

Thanks a lot!

Minoo

On Monday, October 19, 2020 at 3:36:27 PM UTC-4 Agnieszka Dzikiewicz-Krawczyk wrote:

[Agnieszka Dzikiewicz-Krawczyk]

Dear Minoo,

Happy to hear that you succeeded in preparing your libray, either way :) Hope you make some great discoveries with your screens. I was quite frustrated while struggling to get our library cloned, it's just a tool for the proper experiments. But it was worth it :)

Best,

Agnieszka

[Junru Feng]

Hello，Minoo！

Congratulations to you! I am struggling in preparing my own library, I tried CPEC and Gibson assembly, but both the method give me only 7 fold more colonies than control. Can you share your experience with me？ I described the detail in this link, https://groups.google.com/g/crispr/c/AWgFyTjdMpg.

Hope to hear from you!

Junru 

[Minoo karimi]

I think you are struggling with too much background plasmid. In my case, I realized that ESP3I dose not work 100% in a short period of time. I experienced that if I increase the incubation time, Enzyme cuts more than expected. so instead of ESP3I digestion, I used PCR reaction to make the linear vector. I used Q5 DNA polymerize with 1ng of background vector (do not use more than 1ng of template in 50ul PCR reaction). After amplification,  I added 1ul of BsmbI-V2  following NEB recommendation(https://international.neb.com/tools-and-resources/usage-guidelines/activity-of-restriction-enzymes-in-pcr-buffers#:~:text=Frequently%2C%20a%20PCR%20product%20must,any%20purification%20of%20the%20DNA.). After that I did PCR clean up. I do not like the gel extraction and believe it affects the Gibson reaction efficiency.  

I used 330 ng of vector + 50ng of custom oligos in 20ul of total Gibson reaction. I incubated the sample for 2 hours at 37C. 

After isopropanol purification, I transformed cells with 100 ng of DNA. I got 6M colonies for 1 electroporation. I also got 130k colonies for my negative control. 

Hope it helps you.

[Junru Feng]

Thank you very much，Minoo！

I tried PCR linearized vector with DpnI digest before, and the template was digested well but it introduced other strange circular plasmid (Lenti-CRISPRv2 with some fragment lost)，so I didn't use it. 

Now, I will follow your suggestion to  linearize the vector by PCR and digest the template with BsmbI-V2!  

For another question, I noticed that you incubated theGibson reaction for 2 hours at 37C , does it work well in your experiment as compared with 1hours at 55C？

Regards,

Junru

[Minoo karimi]

Yes, it does. I contacted NEB about the incubation time and they suggested increasing incubation time to 4 hours proven to improve the efficiency. they also suggested to NOT incubate the reaction overnight. 

In my hand, 2 hours worked better. I can check my notes and let you know how many fold more. 

About Dpn1, if you use Dpn1 in a PCR reaction since you still have the DNA polymerase, it will fill the gap and your plasmid with a shorter length. That is why I did not do Dpn1.

Minoo

[Anthi Demetriadou]

Dear Minoo,

good to hear that you managed to clone your library!

Good luck with your following experiments,

Anthi

[Junru Feng]

Hi! minoo~

 I'm glad to tell you that I constructed my own library successfully !  Thanks for your suggestions, it's really help me a lot!

Best wishes,

Junru

[Minoo karimi]

Hi Junru,

I am very glad to hear your experiments worked well. 

Good luck with the downstream applications :) 

[Konrad Ricke]

Hello experts,

my name is Konrad. I have been trying to clone & amplify a 10,000 guides-containing custom library using the lenticrispr V2 (LCV2, addgene 52961) backbone according to the protocol described by Joung et al., 2017. After the electroporation, I ended up having a very poor library coverage (~1-2 colonies per guide per electroporation, instead of the recommended 500 colonies per guide). I confirmed the presence of the cloned DNA (LCV2 + insert) in these few colonies by sequencing, so the cloning process itself seems to work, yet the coverage of the library appears to be far too low.  Does anyone see at which step I could optimize my amplification strategy?

 

In the following I will summarize what I did for the oligo PCR amplification, the cloning and the electroporation:

1)    I amplified the oligos at 0.04 ng/uL per reaction with the Q5 Polymerase High Fidelity 2x MM (NEB, M0491S) & the respective primers. I used the recommended PCR program with adjusted annealing temperature according to the primer sequence and the NEB TM Calculator: Cycle 1 at 98C for 30s, Cycles 2 to 21 at 98C for 10s, 72C for 10s & 72C for 15s, Cycle 22 at 72C for2 min. I had a clean product in the agarose gel at the expected size (140 bp, no primer dimers) which I purified with the Macherey-Nagel NucleoSpin Gel and PCR Clean up kit (MN 740609.50). This kit should purify DNA fragments down to 50 bp.

2)    I digested the LCV2 backbone (50 ng/uL) with the Esp3I (BsmBI, 1 ul per reaction), together with alkaline phosphatase (1 uL per reaction) & 1 mM DTT at 37C for 60 min. The Digest was confirmed by running the DNA on a gel (12 and 1.5 kbp fragments), followed by the purification (again MN 740609.50) of the 12 kbp fragment.

3)    I set up the Gibson assembly reaction by using 2x Gibson MM (NEB E261S as well as a custom-made 2x MM from a nearby lab as 2 separate cloning runs, both leading to similar low efficiencies), 330 ng (16.5 ng/uL) of the digested, purified LCV2 backbone (12 kbp) and 50 ng (2.5 ng/uL) of the amplified, purified inserts or H20 as a negative control for the Gibson. I incubated the reaction at 50C for 60 min, followed by Isopropanol purification of the DNA.

4)    I electroporated the DNA at 100 ng/ul with 2.5 uL in 50 uL of electrocompetent cells (Lucigen cat # 60242-2) in a 0.1 cm cuvette. I also EPed DNA of a purchased library in the LCV2 backbone as a positive control for the electroporation, which gave me a bacterial lawn with the non-diluted cells and the expected high coverage on plates with diluted cells. The number of colonies of my custom library was very similar to the negative ctrl of the reaction, however (as mentioned above) we confirmed presence of the cloned product by sequencing.

 

I also tried the CPEC cloning strategy with the same backbone and library which didn’t work in my hands.

 

It would be great if you would have constructive suggestions.

 

Thanks for your help.

 

-Konrad

[Anthi Demetriadou]

Dear Julia,

I have some rather general questions.

1. Regarding representation in the stage of cell transduction, you recommend to aim for a coverage of >500 cells expressing each sgRNA. What is the minimum suggested representation? Does a representation of ~250x consider very low? How a low representation can affect the experiment? (my library is very small, 1250 sgRNAs).

2. In the case of a FACS screen, as shown in the question above, you recommended to subject in NGS the plasmid DNA library, the top 10% fluorescent cells, the bottom 10% fluorescent cells (sorted control) and the diffentiated but not sorted cells (unsorted control). 

I was wondering, since I have already performed NGS of my custom plasmid library to assess sgRNA distribution and confirm full representation, is it necessary to include the plasmid library in the NGS that will be performed for the screening samples? if yes, what will be the purpose of including it?

Also, I'm not sure if I understand the usefulness of analyzing an unsorted control sample with NGS. Is it necessary to include it? From what I understand, the samples required for analysis using MAGeCK are the top 10% and the bottom 10% fluorescent cells. Is that correct?

Thank you for always answering our questions!

Best wishes,

Anthi

[Matthew Pandelakis]

Hi all, 

I too am having some troubles with low assembly efficiencies using Gibson. Looking at the reaction set-up suggested in Julia's paper, it is recommended 330ng of LV2 and 50ng of amplified library. If i understand correctly this is a 15:1 insert to vector ratio which seems very high, is this necessary? 

Additionally, NEB's Gibson Assembly protocol says that is optimized for 50-100ng of vector. Why so much vector backbone? 

Has anyone tried just 100ng of vector and an insert vector ratio of 2:1?

Thanks, 

Matt