Protocol Online logo
Top : New Forum Archives (2009-): : PCR, RT-PCR and Real-Time PCR

Reverse transcription 101 question! - (Oct/29/2013 )

Dear all,


I have a reverse transcription 101 question!  Does anyone know what happens to the cDNA/RNA hybrid molecules during a reverse transcription reaction?  What I'm asking is, after a reverse transcriptase (RTase) lands on a mRNA molecule and synthesize the first-strand cDNA molecule, does the cDNA/RNA hybrid molecule stay as a hybrid until the end (i.e. it's impossible for that mRNA molecule be used as a template again), or does the cDNA/RNA hybrid molecule simply fall apart and the mRNA is available as a template to make more cDNA molecules?


The reason I ask this is because I'm trying to develop a digital PCR assay that would allow me to quantitate the amount of mRNA in my samples in an absolute quantification manner.  I got the digital PCR part pretty well validated already.  However, unless there's a cDNA synthesis system that would trully copy 1 RNA molecule into 1 cDNA molecule (and not more), then there's really no way to quantitate the absolute copy number of mRNA molecules...


Thanks for your help!  I appreciate any thoughts and advice on this!




We use RNase H to degrade the RNA hybridized to cDNA. So it does stay hybrid.


Hi Curtis,


Thanks for the note, what you said is what I have believed in my whole life too, until I spoke to an "RNA specialist" recently, who claimed that such cDNA/RNA hybrids don't really exist, and that the RT reaction is linear with respect to factors like primer concentration... etc.  He believes that the cDNA/RNA hybrids aren't stable and simply fall apart readily, such that the mRNA molecules are available for another RTase immediately.  I'm now very confused, but I can't seem to find any publications that support these theories one way or another. 



There is no reason to do anything to the hybrid if you are going to run a PCR since it will seperate when heated.  THere would be no way to do an absolute quantification of the RNA product unless you  could tag the mRNAs and know you gone all of them.


Hello Kuma,


I cannot say with absolute authority but I'll add some thoughts based on our experience.


1.  RNA/DNA hybrids are actually quite stable - more so than DNA/DNA hybrids actually.

2.  The RT enzyme has an inherent RNase H activity.  However, popular RTs like SuperScript have point mutations in their RNaseH domains to "reduce" RNase activity.  Note they do not say it is totally eliminated.

3.  It appears that the RT reaction does indeed create a small amount of amplification - suggesting that some templates are read more than once - but how much this occurs is hard to determine. IE if you take 10 ng of mRNA - and do an RT - you may get 12 ng of cDNA.  But how accurate this is is really hard to determine and it is not consistent in my experience.

4.  I am sure some templates get read more than once - but how much is hard to say.

5.  As Curtis points out, for many protocols, after cDNA conversion - there is an RNase or heat degredation step to remove the RNA.  Now is this really to break up the hybrids?  Really we do it to make sure there is no RNA left which can interfere with down stream hybridisations on a microarray for example.


If you really want to know the absolute number (not just a really close estimate - which I think you will get even if there is some duplicate reads of a few RNA templates) - do you have anyone nearby that can run a NanoString assay?  Theoretically the NanoString assay should give you a direct count - as it simply relies on tagging the RNA in solution.  There is no cDNA conversion.  It is quantitative and very robust.






I've heard several disputes by companies, that either degradation of RNA:DNA hybrids is necessary for precise qPCR (usually by RNase H+ RT enzymes manufacturers) and that it doesn't really matter much, and RNase H is degrading RNA even before transcription, so reducing RNase H activity makes much higher yield (usually by SuperScript manufacturer).


I usually stick with the RNase H+ RTs for qPCR, because it makes sense to me more. RNA presence can presumally inhibit PCR, question is how much of it is left there after the denaturation.


However I think that the conditions of the mRNA before transcription, isolation procedure, fragmentation, types of RT primers use affect the final 'precise' number much more than the questionable unlinearity of RT enzyme. Unless you are perfectly sure, that the mRNA is the same size, and same condition (and, you generally can't), I wouldn't worry about RT linearity of common 'qPCR suited' RT.

Of course, those with RNase H can maybe decrease the "true" number of molecules, and those RNase H-reduced may increase the "true" number, you can flip a coin (or alternatively try both and compare) but mainly I don't think you can get the "real true" number you wish by any RT-dPCR modification. There will be AFAIK still a certain inaccuracy, higher than some +/- 10 molecules.