I've lost my introns, can you help me find them?

April 5, 2006 12:51 PM

Calling all biologists - I've amplified a 700 bp region of cDNA for the gene PGI and now am trying to amplify this region using genomic DNA. Way more detail inside...

Relevant background: PGI in plants appears to have 21 introns found in conserved locations, throughout the plant kingdom. My goal is to design primers to amplify some of these introns.

So far I have amplified this 700bp region in cDNA using degenerate primers. With the sequence I got from the cDNA I designed species specific primers around the intron boundaries. Using these species specific primers I have tried amplifying this region using genomic DNA. The first month of amplifications each time I amplified the genomic DNA I would run the product out on a gel and every sucessful amplification would show a smeared band the same size as was found in the cDNA. I was also having a bit of an issue with negative controls during this time and both my advisor and I were convinced I had contaminated something with cDNA. I threw out all of my reagents and primers and started all over again designing new primers based on the cDNA sequence data. Using genomic DNA that had never been opened at the same time as cDNA was being using I ran a gradient pcr reaction with my new primers. Yet again I am getting bands the same size as my cDNA, but this time they are not smeary at all, and further, my negative controls are perfectly clear. I've picked the brains of all of my labmates and haven't come up with any reasonable solution and my advisor is out of town for another two weeks. Does anyone have an idea about what could be going on? Could I be amplifying mRNA (I can't imagine it wouldn't have degraded in the DNA isolation process)? Has anyone ever heard of a gene losing all of its introns (losing 21 seems virtually impossible)? Can anyone think of some experiments I could do to identify what I am amplifying?

posted by a22lamia to science & nature (25 comments total)

Do you have any positive controls using the genomic DNA, for example of other regions from the same cDNA?
posted by exogenous at 12:59 PM on April 5, 2006

I haven't been able to amplify any other regions of this gene, so for lack of a better positive control I use an intergenic region in the chloroplast.

A little more detail on this last round of amplification: I used three different 5' primers one aat 805bp, the next at 902bp and the final at 1025bp. I used the same 3' primer with each reaction found at 1305bp. Obviously the bp position is based on the entire gene sequence available from other plant species.
posted by a22lamia at 1:08 PM on April 5, 2006

What plant are you trying to amplify in? Could you switch to a different organism and see if the same thing has happened?

To clarify, did the sequence from degenerate primers reveal the presence of those 21 introns in this particular organism? Could you try amplifying just a single intron (if any are sufficiently large to see on a gel)?

Are there any known chromosomal DNA sequences in the plant you're using as template? Could you try to amplify some of those?

And in those 3 different reactions you just mentioned, do those all yield a product of the exact same size you'd predict from cDNA?
posted by rxrfrx at 1:54 PM on April 5, 2006

I can't switch to a different organism because the primers are specific to my species, or at the very least the genera I'm working with. The sequence from the degenerate primers was amplified from cDNA which only contains the coding domain of the sequence so it wouldn't reveal the presence of introns. The primers I designed are around known intron boundaries within plants and include between 2 and 7 introns, most introns within PGI are relatively small and would be very hard to see on a gel if I focused on only one. The products (and sequences) of the genomic PCRs are identical to the products of the cDNA PCRs.

I'm appreciating all of these questions, they are giving me different views about how to look at these data and love to get more of them, thanks for the help so far!
posted by a22lamia at 2:14 PM on April 5, 2006

Starting from the obvious:

the 700bp PCR fragment sequences out to be the gene you're actually interested in, right?

Have you considered making a different primer set for the genomic amplification than for you cDNA amplification?

Stupid question - have you BLASTed your primers for specificity?

How long is the entire genomic sequence? If you're approaching 1 or 2 kb (or more) consider using a more expensive polymerase like Platinum Taq.

Do you RNAse your genomic DNA preparations?

For a positive control, is the genomic gene (with introns) sequenced? Make primers that will bind to the introns...

What do you want to do with the amplified genomic fragment? Subcloning? Manipulation?

If you're getting a smeared band (that's big enough to be your genomic PCR fragment), you could electrophorese the band onto nitrocellulose and do a blot against a fragment of your 700bp PCR product...

Sorry if these are really obvious suggestions. Just covering bases.
posted by PurplePorpoise at 2:59 PM on April 5, 2006

Am I correct in assuming that the intron sequences are not conserved? If they were, you could design a primer with intron sequence, or one that overlaps the intron/exon boundary with its 3' end in the intron.

Have you tried using primers that straddle the predicted splice sites? Those would be predicted to only make a product if the introns are indeed absent in the template.

You say that the genomic DNA had never been opened whilst cDNA was being used - that sounds like it's a different aliquot of the same prep. Do you have material available to make a fresh DNA prep under controlled conditions, adding RNAse each step of the way or whatever to make sure there's no cDNA or mRNA present?

Whilst I'm struggling to think of an example, it is not implausible that during evolution a spliced mRNA could find itself being reinserted into the genome with the help of a friendly retrovirus or something. It would be rotten luck for you though. In your degenerate PCR experiment, did you recover more than one clone with different sequences? You might find that your plant has a PGI with introns and a pseudogene without or something horrid like that.
posted by nowonmai at 3:13 PM on April 5, 2006

Hi PurlplePorpoise, Thanks for the questions!,

The 700bp fragment blasts as PGI, the primers are new, based on the cDNA sequence and are specific for my genus, although I haven't blasted them because they are so short (21 bp each). The entire genomic sequence is about 6,000bp, I'm dealing with about 1/3 of that region broken down into even smaller sections. I've used FastTaq and Platnum Taq for amplification. I RNAse all of my genomic preps. The genomic gene is not sequenced, I'm hoping to seq some of the introns for use in a population genetics study so I can't make primers that will bind with the introns. The bands I'm currently getting are very clear - no more smearing.

I'm super glad you are covering all the bases both of my advisors have been away off and on the last few weeks - I'm welcoming all of these checks on what I've done
posted by a22lamia at 3:34 PM on April 5, 2006

And now for the answers to nowonmai's questions:

While the intron locations are conserved, the sequences themselves are not.

I don't think using primers that straddle the predicted splice sites would give me anymore information as I've sequenced the bands that I've gotten thus far and they all match the cDNA sequences >99%

The cDNA was synthsized from RNA isolated from a living specimen in my lab while the genomic DNA was isolated from a sample collected in the field. I could definitely repeat the genomic isolating using RNase at each step - I'll try that tomorrow.

As far as the degenerate PCR experiment I direct sequnced the pcr product so there was no cloning involved. Looking at the amino acid sequence though it is nearly identical to other functional PGI genes from other species so I don't think it is a pseudogene.

Thank you for the questions/advice let me know if any more come to you!
posted by a22lamia at 3:46 PM on April 5, 2006

BLAST those primers!!!

There may be sequences elsewhere on the genome that's close enough that your primers will bind to those sites.

Do your primers hairpin? The potential for primer-dimers?

All of my on-line tools bookmarks are on my other computer. If you want, post the primer sequence and I'll eyeball it.

Do you have access to someone who's really good at sequencing dirty DNA? You could design sequencing primers at the edges of the cDNA that sequences OUT - use those primers on genomic DNA and hope you get good-enough-quality signal so you have more sequence with which to design good primers to sequence INwards...

Also, in the covering-al-bases vein, you're using aerosol-resistant tips, right? Have you tried using a different set of pipettes/someone else's PCR plasticware?

With RNAse - don't RNAse for too long, there's *some* cross-specificity with DNA and it might just bork your template enough to give you poor-quality product.

Hmm, have you tried using betaine or DMSO on the genomic template? One problem with genomic DNA is that it's all tangled and twisted up. Try using between 1% to 10% DMSO in your final reaction volume. I find that people generally only bother to try betaine when the DNA's really GC rich or as a last resort.

How are you extracting your genomic DNA? Is there a more "gentle" method (ie., embed the plant cells in a LMW Agarose plug, let the lysozyme diffuse into the plug and work on the cell walls, rinse well (let it difuse back out), let proteinase diffuse into the plug, let it work, rinse well (let it diffuse back out), and either agarase the plug or electrophorese the DNA out of there...
posted by PurplePorpoise at 6:20 PM on April 5, 2006

Here are my primer seqs:
FORWARD
805 5'- GATGGGTTGGTGGTCGCTAT
902 5'- GAGCTCCAGCATTGACAATCA
1025 5' - CATACTCTCAGGCGCTAGAGAAG
REVERSE
1305 5'CATCGTGGTTGCTCACAATCT

I can't see any areas that would hairpin, but an extra set of eyes always helps.

And I agree, blast those damn primers... Oh you mean blast them. I'm in the process of doing it now - Can you give me an idea of what thes can tell me? So having blasted them it appears that 805 is the only one that isn't extremely species specific for PGI, but I need a better idea of what you are thinking I can learn by blasting them?

I'll check with the lab that runs our sequencer about the dirty DNA sequencing. My next task is to try using 5' and 3' RACE to get the rest of the sequnce so that may help as well.

Back to the bases, I am usingaerosol-resistant tips. And while I haven't tried a different set of pipettes or plasticware (our lab has sharing issues out of contamination fear) I have dissasembled my pipettes and cleaned them, cleaned and uv'ed all of my plastic wear.

I will look into DMSO - I hadn't even thought about it though, other than getting rid of RNAses what else does it do?

At this point I am extracting my DNA using Qiagen's DNeasy Plant kit. My tissue appears to be bogged down with secondary compounds and it has been the only way that I've gotten high quality DNA that amplifies (at least cpDNA) consisently.

Thank you again!
posted by a22lamia at 7:22 PM on April 5, 2006

Wow, that's really odd, getting the same (spliced) sequence from genomic DNA... It would be interesting if it had something to do with the phenomenon of mutant Arabidopsis plants reverting to wildtype, but that's probably a bit too exotic an explanation.

Exactly how small are the introns in other species?
posted by greatgefilte at 7:40 PM on April 5, 2006

Some other suggestions:

-- Definitely get some fresh genomic DNA.

-- When you made the cDNA, what kind of primers did you use? Oligo-dT, random hexamers, etc? If you used oligo-dT, it might be interesting to try random hexamers and see if you can amplify the region from pre-mRNA (that is, before it's spliced, when the introns are still in).
posted by greatgefilte at 7:56 PM on April 5, 2006

I'm going to be a curmudgeon... please don't take it personally.

805: the prevalence of di-G's may decrease specificity. The AT on the 3' end isn't as desireable as a GC-rich 3' end. The 5' end is far more flexible - but optimally, a strong and stable contact on the 3' end lends for better amplification.

902: 3' hairpinning (the least desireable part of the primer for this to happen). Again, the AT-rich 3' end.

1025: looks fine

1305: weak 3' end. Primer dimer w/ 902.

The DMSO isn't so much getting rid of proteins, but it's useful for "opening up" DNA - GC rich DNA, especially, tends to have significant secondary structure. DMSO helps break some of that secondary structure up. Other sequences (especially sequences with modifiers such as methylation) may have built-in secondary structure, especially if it's a segment that's involved in gene regulation. Intron-dense regions are sometimes problematic - and the DMSO can be useful here.

I've had some experience extracting DNA from plants - not particularly fun. My problem were the phenolics interfering with enzymes and munging up the DNA.

I've had similar problems extracting high quality DNA from soil - hence the necessity for the agarose plug rigamorole.

As for BLASTing (heh) the primers: other parts of the genome may have a similar enough sequence such that your primers may not bind ONLY to the spot you want them to bind. Since your plant isn't completely sequenced, it's very possible that your primers have cross-specificity to other parts of the genome but BLAST won't be able to tell you this.

Given that, it might not be a bad idea to make another 'reverse' sequence and add that to the matrix.

Good luck!
posted by PurplePorpoise at 8:17 PM on April 5, 2006

Oh, another point. Are you using Gilson pipettes (or another brand that you can take apart completely)?

Wash them with warm soap and water. Spray down with 90% EtOH, then soak them overnight in 10% bleach. Rinse well with filtered water.

Also - how are you drying pellets? Perhaps try airdrying them in a different part of-/lab (I've seen people aim a fan into their eppendorfs - don't do this). If you're using a Speedvac (centrifuge + vacuum), perhaps try just airdrying instead (I've seen a problem wherre someone forgot to open their eppendorf and it ended up expoding inside the speedvac - the PCR product ended up contaminating everyone else's open eppendorfs).

The postdoc I was working with 2 years ago had a similar problem to yours - we could never figure out what was going wrong. His primers were from another lab and they swore that they worked and werre specific. The solution ended up being me making new primers to an entirely different part of the gene and running good negative controls.
posted by PurplePorpoise at 8:24 PM on April 5, 2006

Thank you for eyeballing the primers - I didn't catch the 902 hairpin+ primer dimer. Because of the rediculous amount of introns in this region its hard to find good areas to make primers, but as primers are pretty cheap I think I'll design some more today.

There is someone in my lab working on soil who is having some trouble extracting DNA - I'll pass on the agarose plug info.

Thanks for all of your advice - I'll let you know if anything interesting happens!
posted by a22lamia at 4:48 AM on April 6, 2006

As a non-biologist, this was the most entertaining conversation I've read where I understood absolutely nothing :) However, I have a desire to gain a little knowledge in this area.

What's a good starting text or two for an armchair DNA mechanic? ;)

Apologies for derail!
posted by lowlife at 5:16 AM on April 6, 2006

lowlife, the standard manual for common molecular bio procedures is Sambrook, but you'd probably need an intro bio class to understand what you're reading.


A good introductory molecular bio text is Genomes, available for free online in its entirety.
posted by rxrfrx at 6:48 AM on April 6, 2006

a22lamia: I agree with PP's ideas about trying to generally optimize your PCR and possibly ignore the aberrant band (that likely only looks like the size of your cDNA, but isn't).

Things I've found helpful in plant PCR:

Try a few different protocols/kits for DNA extraction

Oxidized phenolics will mess up your PCR... you can try to deal with this by adding some solid PVP (a tiny pinch) to your plant tissue while grinding in lN₂

In addition to DMSO, try betaine or "Perfect Match" (thermostable SSB) to remove secondary structure

Run annealing temperature gradients over a huge range, not just in the range you think makes sense (e.g. 40-70 °C)

posted by rxrfrx at 6:53 AM on April 6, 2006

I'd have to agree with the chorus that optimizing the primers would be a good idea. I ran them through PrimerQuest with a dummy sequence to amplify, and they came back with penalty scores of 17-22, which in my experience are pretty bad. I've had a lot more success with that program than with Primer3 (upon which PrimerQuest is actually based, but that's another story). If you can get primers to have a penalty score under 5, that's fantastic.
posted by greatgefilte at 7:56 AM on April 6, 2006

Sorry, the link to PrimerQuest is this: http://scitools.idtdna.com/biotools/primer_quest/primer_quest.asp.
posted by greatgefilte at 7:57 AM on April 6, 2006

Thanks for the additional advice, I'm really constrained on primer design, but I'm going to go back through and see what I can do.

rxfrx, I've sequnced this band and it is the exact sequence I have gotten from my cDNA so it isn't just an erroneous band. I will however take your DMSO, Betaine, etc and Gradient advice.

Thank you so much -
posted by a22lamia at 9:22 AM on April 6, 2006

You're welcome... personally, although it seems unlikely, I'd guess it would be possible that there really is an intronless version of this gene in your plant.
posted by rxrfrx at 9:38 AM on April 6, 2006

Regarding extracting DNA from soil - what is the target organism?

I eventually had to give up directly extracting from soil (although in retrospect, adapting lN₂ techniques from plant extraction might have improved DNA quality) and ended up doing a hydrophilic/hydrophobic extraction with mineral oil for actinomycete spores - washed the spores, then went after the DNA.

rxrfrx has a point - do you have any sequence of any of the introns? A good control would be to have one of your primers to anneal inside an intron. If you get product, then you'll know that it's from where you think it's from (as opposed to amplifying an intronless copy of the gene).
posted by PurplePorpoise at 9:51 AM on April 6, 2006

My labmate works with ciliates I think he may have to go for something along this line.

I only have sequences of the introns in other species, none of which are closely related and none of which seem to have areas of conservation.

I've been working with the IDT link you gave me and as luck would have it most of the primers it has designed haven't crossed introns!
posted by a22lamia at 10:11 AM on April 6, 2006

a22 -- If you go the 'advanced' tab in PrimerQuest, you can specify what the target sequences are and where the primers should be (or shouldn't be, if you want to exclude exon-exon boundaries). Try playing around and see what you get!
posted by greatgefilte at 11:30 AM on April 6, 2006