genotyping by PCR

Model Organism: Worms Scientific Discussion
1

Hi everyone. I’ve been performing screening assays by PCR for genotyping worms. Everything was OK except that know I have no amplicon in the PCR,
altough positive and negative controls are OK. I’m now wondering if the DNA extraction failed. Did any of you ever have the same problem?
Dou you have any checked protocol for genomic DNA extraction o have any advices?
Thanks!!

2

Here’s the protocol we use, which is very robust:

  • Pick 10-12 worms into 2.5 uL of lysis buffer (10 mM Tris pH 8.3, 50 mM KCl, 2.5 mM MgCl2, 0.45% TX-100, 0.45% Tween 20, 0.01% Gelatin, 0.1 mg/mL Proteinase K) in a PCR tube. You can also do single worms, but it’s less robust. A better approach to genotyping single worms is to single, let the worms produce progeny, and genoptype 10-12 progeny (if you pick 10 worms, you have a >99.999% chance of seeing both alleles in the offspring of a heterozygote).
  • Put at -80C for 5-10 min until frozen, then move to thermocycler. Incubate at 65ºC for 1h, then at 95ºC for 15’ to inactivate Proteinase K.
  • Add 22.5 uL of PCR master mix (we use NEB’s LongAmp Taq) and run 30 cycles of PCR. Make sure you calculate the annealing temperature with the calculator on NEB’s website, which takes into account the composition of the LongAmp buffer.
  • If you’re doing RFLP, add 25 uL of restriction enzyme mix directly to the PCR tubes and digest for 1-2h before running on a gel. Many restriction enzymes will work in 50% LongAmp buffer / 50% digest buffer, but check the compatibility here: https://www.neb.com/tools-and-resources/usage-guidelines/activity-of-restriction-enzymes-in-a-q5-taq-or-phusion-pcr-mix.
3

Dan is providing a very standard protocol; I’d only offer a couple of suggestions for things you might consider:

  1. Dan’s lysis volume (2.5 ul) is very small. Evaporation can be a concern, unless you’re putting mineral oil over your lysis reactions (which you can do, but isn’t worthwhile). This might be somewhat different for different PCR machines. I typically do the lysis in 10-12 ul for this reason.
  2. Protocols typically advise to transfer minimal bacteria into the lysis, which you can do by picking the worm to a region of the plate off the lawn, flaming your pick, and scooping the worm up for transfer to lysis buffer. I don’t know if this actually makes a difference, though.
  3. Lysate is not stable and should be used shortly (within an hour or so) after heat-inactivation. It doesn’t store well, even at -80.
  4. I’ve found that lysate from a single adult can extremely reliably (>95%) be used to PCR a half-dozen markers, including heterozygosity; others have I think reported extremely reliable PCRing of tenfold more (~60) markers from single worms. Because of this I dilute my lysate several-fold with water before setting up reactions that each use only part of the lysate; this makes pipetting easier and by diluting any schmutz in the lysate and diluting the lysate buffer may make PCR more efficient.
  5. I would disagree with Dan a bit about single worms versus progeny: because I’ve found single-worm PCR to be extremely robust, I feel it can be easier to target one large worm instead of several small ones, there are no issues about getting a representative subset of the progeny, and there are generation-control issues if you’re worried about losing a heterozygote that can reward genotyping an adult while its progeny left on the plate are still young. On the downside, finding the single adult can also be harder than scooping up a few of the many progeny, and the adult has to be genotyped a day or two after being left on the plate, which might not be a convenient time.
  6. Since Dan is offering his reaction conditions: my standard conditions were, as I recall: 2.5 ul oligos (9 pmol each), 10 ul diluted lysate (diluted at least 1:2 with water), 10 ul PCR master mix (Roche Taq buffer, slightly elevated dNTPs, 0.5 U Taq). PCR for 35 cycles, typically annealing at 55C, using oligos usually found by Primer3 at default settings. Product size usually 400-600 bp, with melting, annealing, and extension times all 30 seconds or less. When doing restriction digests, I’d add 5.5 ul digestion mix that typically contained 2.8 ul 10x NEB reaction buffer and 0.1-0.2 ul enzyme; I don’t recall this ever not working (obviously, do a control first).

Good luck!

4

Those are all good points. Just quick replies for a couple of them:

  1. Lysis volume: PCR reactions can fail if the lysate makes up too large a volume fraction (>10%) of the PCR reaction. Keeping the lysate volume small allows direct addition of PCR master mix to the tube in which the lysis was done, which eliminates extra dilution/pipetting steps. This is just a matter of personal preference.

  2. Lysate stability: It is true that the lysate is not stable after proteinase K digestion and inactivation. However, in my hands the frozen worms (during the freeze-thaw step, immediately after picking and before incubation at 65C) are good for at least a couple of weeks. Once I’ve removed the worms from -80 and done the proteinase K digestion, I always use the lysate right away.

  3. Worms vs. progeny: I should have said this in my earlier post: it depends on what size amplicon you’re after. For ~500 bp, single worms are fine. But when I use this protocol, I am typically trying to amplify 3-8 kb for sequencing (to confirm a CRISPR modification). Getting these larger amplicons from single worms can be really finicky, in my hands. Using multiple worms (and thus more DNA) helps make larger amplicons amplify more consistently.

5
1) Lysis volume: PCR reactions can fail if the lysate makes up too large a volume fraction (>10%) of the PCR reaction. Keeping the lysate volume small allows direct addition of PCR master mix to the tube in which the lysis was done, which eliminates extra dilution/pipetting steps. This is just a matter of personal preference.
I'd certainly agree that the lysate shouldn't be too large a proportion of the final PCR mix - but by lysing in a slightly larger volume I'm more confident of the lysis, and while diluting is an extra step it makes the later pipetting steps easier and makes sure the lysate doesn't dominate the PCR mixture.
5) Worms vs. progeny: I should have said this in my earlier post: it depends on what size amplicon you're after. For ~500 bp, single worms are fine. But when I use this protocol, I am typically trying to amplify 3-8 kb for sequencing (to confirm a CRISPR modification). Getting these larger amplicons from single worms can be really finicky, in my hands. Using multiple worms (and thus more DNA) helps make larger amplicons amplify more consistently.
Fair enough. I've done a fair bit of long-PCR from crude lysate, but always from a bunch of worms. I was thinking of single worms for genotyping known variants, where smaller amplicons are the rule.
6

We’ve done thousands of these. Some with CRISPR (where I suspect Dan has done many, many,many), but most with ok and tm allele deletions, and occasionally PCR of a SNP RFLP.

I’ve found single worm PCR to very, very reliable, given certain approaches:

  1. We keep the amplicon as short as possible. When working under 500 bp, we rarely have any problems. Be well aware of band competition. Longer amplicons will OFTEN disappear when visualizing smaller bands (see triplex, below)
  2. We ALWAYS do triplex PCR for detecting deletions. That way we can reliably visualize m/m, m/+ and +/+ in the same animal (as Hillel described: pick L4 or adult, come back the next day and lyse it after it has laid eggs)
  3. We carefully validate primers before using, with the controls I describe in 4. Sometimes we need to re-design one or more of our triplex primers to get robust amplification without obfuscating ghost bands.
  4. We ALWAYS run controls in parallel for every genotyping experiment: +/+, m/+, m/m. This matters a lot, because band intensity changes based on genotype (see amplicon competition in #1). We keep some of these for years with no loss of efficacy. Think about the hassle of making m/+ animals, and then realize the utility of making excess and storing them for years (at -80, promptly stored after lysis/inactivation), for popular mutations in the lab, or your personal project
  5. We do 2.5 ul of lysis buffer, always prepared fresh, and use oil on top. Prompt lysis, heat inactivation and storage at -80, though overnight at 4˚C post-lysis/inactivation doesn’t seem to harm efficacy or storage longevity. Some hassle from oil in gel loading, but people don’t seem to mind so much after practice to get their gel loading technique down. I’m open to alternatives, but heated lids don’t sufficiently counteract evaopration in 2.5 ul WLB, as Hillel notes.

So I wonder whether Dan’s multiple worms in the lysate come into play with with the longer amplicons necessitated by his bigger homology arms used for his wondrous SEC CRISPR strategy (which we still love, Dan. I keep expecting a little shrine to you to pop in the lab, votive candles and all!) I stress out about our ability to detect the knockins with those longer amplicons, and for one we had to circle back and detect the mutation via nested PCR. Perhaps we could have avoided that with 10-12 animals in the lysate? Worth thinking about.

Another observation: we’ve been doing a lot of gene, or even exon, knockouts lately. The first one was very low efficieivency, but subsequent efforts were quite productive. For these with use the dominant dpy-10(cn64) co-CRISPR marker, and do PCR on pools of 8 Rol F1s, after they have laid their eggs. With the small amplicons (and also triplex) we have no trouble detecting mutations with DNA omplexity of 1 in 16. So maybe those amplicons of <500 bp are the trick for dirty lysates in single worm PCR.

Parenthetically, we’ve done quite well with the knockouts using Seydoux’s ssODN spanning the deletion site (between two guide RNAs). Many of our deletions from this are quite accurate, so that is nice: precision deletions! I like what we’ve gotten from that approach better than what we got allowed ends to bash together. And though anecdotal, I think we’ve gotten a big boost in efficiency with the approach, as her Paix et al paper suggests.

Hope this helps!
Dave