Build gRNA libraries via Gibson Assembly
Updated 4/15/2016 08:38 PM
Introduction
Materials
Procedure
Step A. Design and order gRNA oligos
- Make sure oligos have ~20bp overlaps as follows with your 20bp specificity determining sequence (SDS) in the middle. Use the webtool http://lp2.github.io/yeast-crispri/ to design effective gRNAs for CRISPRi in yeast.
Table1
A A | B B | C C | |
1 1 | Left overlap | 20bp SDS | Right overlap |
2 2 | ctgggagctgcgattggcag | NNNNNNNNNNNNNNNNNNNN | gttttagagctagaaatagc |
Step B. Amplify gRNAs and digest plasmid
Step B-1 and Step B-2 can be done concurrently.
Step B-1. Digest plasmid with NotI
- Digest pRS416gT-Mxi1 with NotI, preferably NotI-HF. You will want to digest a large amount of vector and digest for a long period of time.
We generally digest overnight 8-12 hours or so to reduce background
- Heat inactivate (We generally do extra, 80C for 20 min) after the digest has completed.
DO NOT PCR PURIFY. DO NOT PHOSPHATASE TREAT. These will reduce your efficiency and do little to reduce your background and may even make it higher relative to your successful product.
Step B-2. PCR your oligos
Note, this is not necessary if you are only cloning a single gRNA oligo but not amplifying first will reduce your efficiency
- PCR your library using your sub library specific primers if you have any.
- Use Kapa HiFi Hotstart and follow the directions provided if you want to reduce the GC bias introduce by PCR. Do a large volume PCR (say 80µL) but split it into 4 or more tubes. This helps to reduce PCR jackpotting of any one species in the pool. Do this PCR for 25 cycles or fewer as to not create too much bias.
- Do a PCR with your constant inner primers. For Gibson assembly use primers that increase the length of your homology to the plasmid. I use the following primers which create 40bp overlaps.
3'gRNA-extender gactagccttattttaacttgctatttctagctctaaaac
5'gRNA-extender tcgcggctgggaacgaaactctgggagctgcgattggcag
- I do this PCR at TM 60°C, and for ~30 cycles.
Please note I DO NOT DO PCR CLEANUP. I use the PCR product directly in Gibson assembly.
Step C. Gibson Assembly
- My exact Gibson protocol can be found here: https://benchling.com/protocols/L3HVM4/gibson-assembly. I always use a 4 to one plasmid volume to insert volume ratio. Protocols will tell you to calculate exact ratios but these seem to work well. I attempt to digest my plasmids such that the final digest has a concentration of 20ng/µL. You can reduce the Gibson volume down to as low as 2.5µL total reaction volume for cloning single guides. For libraries though I generally use larger volumes. You only need as much volume as you plan to transform. The detailed protocol, including some of my modifications are available at the benchling link.
Please DO NOT FORGET TO DILUTE THE EXONUCLEASE first while prepping your Gibson Mixes. Do not attempt to pippete exactly .64µL as the original protocol says. I pipette 6.4 of my diluted exo. Using too little or too much exo both cause a reduction in efficiency.
Step D. Transform E. coli
- For small libraries, I used DH5alpha cells I prepped myself using the Zymo Mix & Go EZ competent kit. If you choose to use this kit make sure to grow the cells at between 20-22 °C and in Zymo Broth or they won’t be sufficiently competent. Also it is essential you get the cells at the right OD as stipulated by the kit.
- For projects where the library has complexity greater than about ~10,000 you will either need to transform >100µL of these competent cells OR use electro competent cells. For electrocompetent cells you can increase efficiency further by first ethanol precipitating and concentrating your Gibson Assembly product DNA. Ideally you want enough efficiency to have greater than 10 colonies per gRNA you wish to include in the pool.
- The nice thing about the Mix and Go’s is you can just plate them right away onto LB + carbenicillin plates. If you are doing a library try using larger plates or multiple plate, or selecting directly in liquid LB while and using a small dilution on a single plate to get a colony count estimate.
- Incubate on ice for 5-10 minutes with the Gibson assembly, then do the optional 42 degree 30 second heat shock. Do not do the SOB recovery, just plate directly. Let cells grow overnight. Try not to let them grow more than 24 hours though as if the cells are unevenly distributed eventually the cells in the sparse regions will grow larger colonies than those in dense regions. (Alternatively you can put this transformation directly into liquid LB + carbenicillin media and prep this directly).
- Wash all of the cells off your plate with liquid LB Carb, then miniprep about 500µL (the OD’s of cells washed off is often really really high so don’t do the normal 2-4 mls of culture). We use Qiagen miniprep kit. Sometimes we miniprep multiple tubes so we will have more DNA to transform into yeast.
Step E. Transform yeast with the miniprep DNA using a standard lithium acetate protocol
- To get more transformants you may want to grow more cells. For very large libraries I grow 50mL of cells and use as much plasmid as I can afford to use (up to the 74µL in the protocol). The protocol provided is the one we used. After growth select on –Ura media. You can do this selection on plates or directly in liquid media. We tested liquid media compared to solid media and didn’t find a noticeable difference in relative abundances of gRNAs in the pool.
Comments
Hi Hannah,
How do you dilute the gRNA oligo to use as a template in a PCR in step B-2 (6)?
How do you dilute the gRNA oligo to use as a template in a PCR in step B-2 (6)?
It depends, are you doing a library or a single guide? If you are doing a single guide it really doesn't matter one way or another. If you are doing a library you can dilute but only to the point where you will not influence the complexity of your library. You'll want an amount of template that is at least 10X, preferably higher, representation of your library. Thus if you have a 10,000 member library you'll want >100,000 molecules as template. Depending on who you use to synthesize the gRNA library there will be a varying degree of bias in synthesis. The more bias in library representation there is, the more coverage you are going to want.
I might be missing something but how does this work if the 3' ends from the NotI site aren't complementary to the insert? Is this resolved by polymerase profreading?
This protocol was created by Justin D Smith and edited by Hannah Shen.