Amplification of specific genes from cDNA - (Jan/18/2006 )
I'm working on a structure project involving two specific proteins. So far I've successfully cloned the genes from one yeast strain, and will likely get four more as soon as ATCC makes me some DNA.
My boss(es) want me to also investigate the genes from human, mouse, and possibly cow. For this, they suggest buying cDNAs. There is a company that has the specific cDNA transcript of one of the genes from human, but for everything else, I may have to resort to a whole library of the species in question. I can get the gene sequences by matching up BLAST searches, and I should have no problem designing primers to amplify those specific genes from a library.
The reason I'm posting a question here is because I have zero experience with cDNA - nor with genes from eukaryotes more complex than yeast (so far in my career I've been mainly a bacteria guy), so I'm hoping to benefit from others' experience. I've looked at a couple of commercial sites (including ntomics, biochain, and origene, among others), and the impression I get is that the cDNAs produced are not necessarily complete representations of genomes - the mRNAs may be incomplete, average transcript size tends to be around 1.5 kb, etc.
In a more question-driven format, I'm particularly interested in knowledge regarding the following
1) Can I trust a random library to contain the particular genes I need for a PCR with specific primers?
2) Is there a big difference in general vs. tissue-specific libraries if you're just trying to pull out one (or two) gene products?
3) Are there any caveats I should be aware of before attempting these experiments which the individual companies might not be telling me? (is it possible the library won't contain my construct, is PCR more complicated than usual, do I need to treat it in any special ways...)
4) Is there a better way to generate intronless gene products short of synthesizing the entire gene?
5) Do you have any suggestions for libraries from a specific company or academic entity which have worked for you in the past?
I can help with #2...
I would say tissue specificity makes a difference only if it makes a difference in the expression of your protein
for example, if your POI is expressed in liver but not in gingival tissues, don't get a gingival library 
the mRNA should be more abundant in tissue that actually expresses your POI and therefore the library may be skewed toward the gene you seek
Thanks aimikins, I'll check into that. Not sure if anyone's done a tissue-specific expression profile before, but its certainly something I'll look for before committing to a library.
A followup question on this topic (for aimikins or anyone)
- Origene specifically warns users about amplifying from a library because of the large numbers of PCR cycles used, combined with the inherent fidelity limitations of standard polymerases (even Pfu). Anyone actually had experience with this? Have you had specific trouble getting unmutated genes from a library? Or is Origine just trying to convince everyone to pay for their construct services?
It seems to me you are going about this the hard way. So long as you know the DNA sequence of the proteins of interest, this is not too difficult.
Okay... lets say you want to isolate a gene. Grab some tissue that it is highly expressed in. Get it fresh and snap freeze it. Easy for mouse and cow, less so for human.
Extract the RNA using Trizol or some such reagent.
Get a reverse transcriptase kit (Qiagen sells a good one called Omniscript). Use random hexamers as your primer and set up a second reaction using oligo dT instead of ran hex.
If you know the sequence of your protein of interest, you can whistle up a couple of primers... best make it at lest 4.. two for the 5' end and 2 for the 3' end.
Now go buy a pcr kit (again I will shill for Qiagen and the Proofstart kit.. nice because it has proofreading ability).
Set up pcr using the primers and cDNA. With a little luck you may have just cloned out your gene of interest. Clone it, sequence and pop it into an expression vector to see if it makes your protein.