WTF
April 28, 2010 3:51 PM Subscribe
Scientists: please, help me explain this cloning anomaly.
Okay, I am trying to clone a gene into a pet32 plasmid. I did the restriction digest and the ligation and got colonies. I then sequenced the colonies. This is where the weirdness begins. I think it would be most helpful if I include a little map:
plasmid---GB1 tag---BamH1site---gene_insert---Xho1site---His tag---plasmid
Here is what is weird: the primers I used to amplify the gene insert obviously include the Bam/Xho sites. HOWEVER, when I sequence my plasmid (using standard T7 primers), I see the following:
plasmid---GB1 tag---BamH1site---scrambled DNA---Xho1site---His tag---plasmid
Before I get to the scrambled DNA, I do see the 6 or so nucleotide overhang past the restriction sites (where my primer lies down to amplify the gene).
I blasted the scrambled DNA and it came back as a 100% match for wild-type bacterial DNA. So...recombination, right? But I'm using a standard pet32 plasmid in DH5-alpha cells. This is done all the time. My very lab uses these same two stocks of cells and plasmids with success, all the time. And yes, I did try different tubes of plasmid/cell to ensure I didn't somehow grab some crazy contamination.
So...anyone have any suggestions at all as to what's going on here?
Okay, I am trying to clone a gene into a pet32 plasmid. I did the restriction digest and the ligation and got colonies. I then sequenced the colonies. This is where the weirdness begins. I think it would be most helpful if I include a little map:
plasmid---GB1 tag---BamH1site---gene_insert---Xho1site---His tag---plasmid
Here is what is weird: the primers I used to amplify the gene insert obviously include the Bam/Xho sites. HOWEVER, when I sequence my plasmid (using standard T7 primers), I see the following:
plasmid---GB1 tag---BamH1site---scrambled DNA---Xho1site---His tag---plasmid
Before I get to the scrambled DNA, I do see the 6 or so nucleotide overhang past the restriction sites (where my primer lies down to amplify the gene).
I blasted the scrambled DNA and it came back as a 100% match for wild-type bacterial DNA. So...recombination, right? But I'm using a standard pet32 plasmid in DH5-alpha cells. This is done all the time. My very lab uses these same two stocks of cells and plasmids with success, all the time. And yes, I did try different tubes of plasmid/cell to ensure I didn't somehow grab some crazy contamination.
So...anyone have any suggestions at all as to what's going on here?
Could the polymerase or any of the other ligation reagents or the restriction enzymes that you originally used to prep the plasmid be contaminated?
posted by i love cheese at 4:11 PM on April 28, 2010
posted by i love cheese at 4:11 PM on April 28, 2010
Maybe I'm misunderstanding, are you saying that your primers only overlap the gene of interest by 6 or so nucleotides? That's awfully short, and it's quite likely your primers are annealing elsewhere in the bacterial genome and amplifying that instead of your gene of interest. I'd usually go with a 15-20 nucleotide overlap and increase the annealing temperature.
posted by Durin's Bane at 4:12 PM on April 28, 2010
posted by Durin's Bane at 4:12 PM on April 28, 2010
In addition to making longer primer overhangs, you could try doing nested PCR with the first PCR using primers that complement vector sequence and then the second set adding in the restriction site and complementing the gene.
It does sound like there's some contamination. My general policy is to not share any reagents for PCR: water, dNTPs, pol buffer or polymerase. Also make new primer stocks.
posted by sciencegeek at 4:14 PM on April 28, 2010
It does sound like there's some contamination. My general policy is to not share any reagents for PCR: water, dNTPs, pol buffer or polymerase. Also make new primer stocks.
posted by sciencegeek at 4:14 PM on April 28, 2010
Wait a minute, 6bp overlap? You need to do at least 20 overlap/
posted by sciencegeek at 4:15 PM on April 28, 2010
posted by sciencegeek at 4:15 PM on April 28, 2010
I have had something similar happen - it was almost as though the bacteria didn't "like" the gene. Other genes went in fine, but certain ones didn't, even using the same protocol. I'm seconding increasing the stringency.
posted by Knowyournuts at 4:34 PM on April 28, 2010
posted by Knowyournuts at 4:34 PM on April 28, 2010
My first step with any pcr weirdness is to dump my primer stocks and make them fresh, and I'm paranoid enough that I aliquot and UV irradiate water so I grab a fresh batch of that too. Those are the easy steps I do out of hand to make sure it's not something simple.
Until then, I think we need to wait for you to respond, are you talking about a 6bp overhang of a restriction cutsite? Or are you talking about 6bp of annealing space into your desired gene? If it's the former, then maybe try to get a small sample of a different plasmid and see if you can insert into that one? I've never had it happen to me but others in my department have found that a specific plasmid just isn't good with their gene, presumably due to size or some other weird reason, that might be worth a try too.
posted by dnesan at 5:43 PM on April 28, 2010
Until then, I think we need to wait for you to respond, are you talking about a 6bp overhang of a restriction cutsite? Or are you talking about 6bp of annealing space into your desired gene? If it's the former, then maybe try to get a small sample of a different plasmid and see if you can insert into that one? I've never had it happen to me but others in my department have found that a specific plasmid just isn't good with their gene, presumably due to size or some other weird reason, that might be worth a try too.
posted by dnesan at 5:43 PM on April 28, 2010
You can sequence the pcr product immediately after amplification and subsequent gel isolation by using the same pcr primers that you used for the pcr reaction as sequencing primers. This should give some information as to the identity of the insert prior to insertion in the plasmid. You don't mention your starting template but it is worthchecking out whether it corresponds towhat you believe it is. I would also nth the point about the length of the primer homology at the 3' end should be at least 15-20 bp.
Generic cloning stains (e.g. Dh5a etc) of bacteria occasionaly do odd things to certain sequences. If the problem is reproducible after checking th veracity of the fragment to be cloned into the plasmid, you may wish to consider bacterial strains like stbl3 that are better able to deal with certain sequences.
posted by SueDenim at 6:23 PM on April 28, 2010
Generic cloning stains (e.g. Dh5a etc) of bacteria occasionaly do odd things to certain sequences. If the problem is reproducible after checking th veracity of the fragment to be cloned into the plasmid, you may wish to consider bacterial strains like stbl3 that are better able to deal with certain sequences.
posted by SueDenim at 6:23 PM on April 28, 2010
I agree with what the others are saying--if I understand correctly, it sounds like you amplified something else on the template genome or perhaps some contaminating bacteria's genome. Definitely use nested primers... they will help you sort of zero in on your gene. Design primers 15-25 in length on either size of your gene of interest but about 100 bp away from beginning and end of the gene. Run the PCR, then run a gel on a portion of the PCR. The gel will help you confirm whether or not the size of this newly amplified DNA fragment is what you expect (your gene + about 200 bp). Now, use your second set of primers (similar to the ones you used the first time, but with at least 15 bp overlapping with your gene of interest) to isolate your gene + restriction sites.
You don't say if you ran a gel after your PCR or after your digest, but I always run a portion of my PCR reaction on a gel to confirm that I have the correct fragment. Then I don't waste my time setting up a digest/gel extraction/ligation/transformation if I don't see what I want!
posted by hooper4 at 7:49 PM on April 28, 2010
You don't say if you ran a gel after your PCR or after your digest, but I always run a portion of my PCR reaction on a gel to confirm that I have the correct fragment. Then I don't waste my time setting up a digest/gel extraction/ligation/transformation if I don't see what I want!
posted by hooper4 at 7:49 PM on April 28, 2010
Dumb troubleshooting questions that I'd be asking myself if I ran into this:
-Is this repeatable if you re-run the reaction? If you re-run the reaction with fresh template, dNTPs, primer dilutions, a different polymerase stock, etc. (in case any of it is contaminated with random bacterial DNA for some reason)?
-What's the DNA source for your gene? If it's genomic, have you tried seeing if anything shows up on a BLAST search for your primer sequences? If it's not genomic, is whatever source it's coming from purified? (If it's in another plasmid, have you sequenced that source plasmid to make sure your gene is actually on it, and not the junk bacterial DNA?)
-Have you run a gel on your PCR product to make sure it is the correct length for your gene? (I always do this, to avoid wasting more time on crappy results.) Are there multiple bands, which would indicate insufficiently specific priming? Have you sequenced the product (using your PCR primers)?
-Are your primers right, and typo-free? Double-check the sequences, and try running an in silico PCR (in any of the expensive cloning software - VectorNTI, MacVector, DNAStar, etc. - or in freeware like Serial Cloner) - is your product correct?
-Your phrasing is a little unclear. Which of these two setups accurately reflects your primers:
-Do your primers otherwise follow good primer form? Do they have similar melting temperatures and do they not easily form hairpins or primer-dimers? Use Primer3 to calculate ideal initial primers in a region if you're doing nested PCR; certainly check all primers with IDT's OligoAnalyzer. Do they have a few extra basepairs at the end (the "4nt" in my primer mockups above) to prevent the restriction sites from being at the very end of the PCR product, which some enzymes hate?
-Have you purified your PCR product in any way, before or after digestion?
-Once you've digested your vector, did you gel-purify it too, and compare it to the uncut vector to make sure cutting is happening correctly? (Have you sequenced your vector stocks to be sure they don't have this bacterial sequence in them already?)
Basically, if your primers only have 6bp of homology with your gene or if they're very similar to unwanted sequences elsewhere in the genome, replace them with adequate ones. If they're theoretically OK, do a run where you do analyze the hell out of each step with lots of sequencing and gels. If none of those steps turns up anything useful, you can try redesigning your cloning process with procedures like nested PCR (which will let you pick more ideal primers for the first PCR reaction, and then the precise ones you want for cloning in the second round) to eliminate off-target amplification, gradient PCR to help ID the best reaction conditions, etc., or you can consider switching to other vector and host systems, in case your current system just doesn't like your gene for no good reason.
posted by ubersturm at 11:49 PM on April 28, 2010
-Is this repeatable if you re-run the reaction? If you re-run the reaction with fresh template, dNTPs, primer dilutions, a different polymerase stock, etc. (in case any of it is contaminated with random bacterial DNA for some reason)?
-What's the DNA source for your gene? If it's genomic, have you tried seeing if anything shows up on a BLAST search for your primer sequences? If it's not genomic, is whatever source it's coming from purified? (If it's in another plasmid, have you sequenced that source plasmid to make sure your gene is actually on it, and not the junk bacterial DNA?)
-Have you run a gel on your PCR product to make sure it is the correct length for your gene? (I always do this, to avoid wasting more time on crappy results.) Are there multiple bands, which would indicate insufficiently specific priming? Have you sequenced the product (using your PCR primers)?
-Are your primers right, and typo-free? Double-check the sequences, and try running an in silico PCR (in any of the expensive cloning software - VectorNTI, MacVector, DNAStar, etc. - or in freeware like Serial Cloner) - is your product correct?
-Your phrasing is a little unclear. Which of these two setups accurately reflects your primers:
F: 5' (4nt) ---- BamHI ---- 6bp linker region ---- 18+bp homologous to geneThe first is likely to be fine - it's pretty standard. The latter is almost guaranteed to give you crappy results, since it's nowhere near specific enough. As a rule, make the bits of your primer that are homologous to your gene at least 18bp long. Much less like that, and it's statistically too likely for comfort that your primer sequence may appear elsewhere in genome.
R: 18+bp homologous to gene ---- 6bp linker region ----- XhoI ----- (4nt) 5'
or
F: 5' (4nt) ---- BamHI ---- 6bp homologous to gene
R: 6bp homologous to gene ---- XhoI -----(4nt) 5'
-Do your primers otherwise follow good primer form? Do they have similar melting temperatures and do they not easily form hairpins or primer-dimers? Use Primer3 to calculate ideal initial primers in a region if you're doing nested PCR; certainly check all primers with IDT's OligoAnalyzer. Do they have a few extra basepairs at the end (the "4nt" in my primer mockups above) to prevent the restriction sites from being at the very end of the PCR product, which some enzymes hate?
-Have you purified your PCR product in any way, before or after digestion?
-Once you've digested your vector, did you gel-purify it too, and compare it to the uncut vector to make sure cutting is happening correctly? (Have you sequenced your vector stocks to be sure they don't have this bacterial sequence in them already?)
Basically, if your primers only have 6bp of homology with your gene or if they're very similar to unwanted sequences elsewhere in the genome, replace them with adequate ones. If they're theoretically OK, do a run where you do analyze the hell out of each step with lots of sequencing and gels. If none of those steps turns up anything useful, you can try redesigning your cloning process with procedures like nested PCR (which will let you pick more ideal primers for the first PCR reaction, and then the precise ones you want for cloning in the second round) to eliminate off-target amplification, gradient PCR to help ID the best reaction conditions, etc., or you can consider switching to other vector and host systems, in case your current system just doesn't like your gene for no good reason.
posted by ubersturm at 11:49 PM on April 28, 2010
Assuming your primers are correctly designed as suggested above (> 18bp of amplicon overlap), and your template checks out ok, you may want to verify that your primers are purified by the vendor to be the correct length- even if this is the case, you may have nuclease or dna contamination in your primers which can lead to odd amplifications. Do a negative PCR amplification control with the primers alone (followed by gel run) which will verify that there is no contaminating template in the primers and that the primers are not forming concatemers. If you suspect the primers are not full length, you can also simulate this by chopping bases off the primer sequences and seeing where they lie on the bacterial genome using Primer-BLAST vs. the bacterial genome you got as an insert. Most primer vendors will replace the order if you are a regular buyer and are getting odd results.
Disclaimer: It's been a while (years) since I ordered primers, so perhaps all vendors purify their oligos these days using HPLC or PAGE as a matter of course.
posted by benzenedream at 1:01 AM on April 29, 2010
Disclaimer: It's been a while (years) since I ordered primers, so perhaps all vendors purify their oligos these days using HPLC or PAGE as a matter of course.
posted by benzenedream at 1:01 AM on April 29, 2010
Response by poster: Sorry, I was in a rush typing. My primers have an 18 nt overhang into the gene (6 amino acids).
posted by sickinthehead at 7:29 AM on April 29, 2010
posted by sickinthehead at 7:29 AM on April 29, 2010
DH5a are certainly capable of scrambling your DNA. I work with a lot of retroviral contructs that contain repeated sequences and had enormous troubles (scrambled sequences and large deletions) trying to clone certain sequences until I moved to using Stbl2 bacteria from Invitrogen. (And these sequences had, I thought, trivial differences from what I'd had success with before - like swapping a myc-tag with a 3xFLAG-tag.) I've also heard of people having better luck using Sure 2 bacteria from Stratagene for some difficult to clone herpesviral sequences.
I'd second sequencing your PCR product before trying to clone with it to see if it is a problem with your PCR or if the bacteria are screwing you over and if it is the bacteria, look for a strain that is more than just RecA-.
posted by The Bishop of Turkey at 6:18 AM on April 30, 2010
I'd second sequencing your PCR product before trying to clone with it to see if it is a problem with your PCR or if the bacteria are screwing you over and if it is the bacteria, look for a strain that is more than just RecA-.
posted by The Bishop of Turkey at 6:18 AM on April 30, 2010
Sorry, I was in a rush typing. My primers have an 18 nt overhang into the gene (6 amino acids).
That should work just fine.
We could be more helpful if you sequenced your PCR product. Is the scrambled region more than 50bp larger or smaller than what you intended? If so if you just ran it on a gel we'd know a lot more. You might need to optimize your PCR or get your bacteria to behave.
Did you mention the model organism or strain?
posted by Blasdelb at 6:23 PM on April 30, 2010
That should work just fine.
We could be more helpful if you sequenced your PCR product. Is the scrambled region more than 50bp larger or smaller than what you intended? If so if you just ran it on a gel we'd know a lot more. You might need to optimize your PCR or get your bacteria to behave.
Did you mention the model organism or strain?
posted by Blasdelb at 6:23 PM on April 30, 2010
This thread is closed to new comments.
posted by be11e at 4:11 PM on April 28, 2010