After talking with two Bio-Rad reps we still don't know what's going on.

They recommended a more conservative range of 10-100ng/20ul reaction (might solve the poisoning or overloading). However, I really don't think it's an overloading issue. Many people claim to load 1ug/20ul reaction and get good results, and this modification would increase the effective price per ng of cDNA by 10x compared to my current protocol.

My purification protocol need to be improved for a variety of reasons: genomic DNA contamination (see DNAse test below), possible degradation (detected on Agilent 2100), whatever this inhibitor is (previous post), etc.

I may try altering my RNA purification to see if this ameliorates the issue. After speaking with a colleague who cares about these sorts of controls, I was given a few recommendations. I more-or-less follow the TRIzol protocol, they recommended changes at these steps. Does anyone else have strong suggestions?

1) TRIzol Ratio: 0.45mg tissue  / 0.5ml TRIzol

• CHANGE: Decrease this ratio by using more TRIzol, may help decrease gDNA contamination.

2) Lysis assistance: Sonication

• CHANGE: Omit sonication. Shearing of gDNA may be making the gDNA contamination worse.

3) Phase separation: Phase lock tubes          

• CHANGE: Omit phase lock tubes, and manually pipette to ensure we don’t collect the interphase. (Does anyone know where the interphase ends up in phase lock tubes? On the TRIzol side, the aqueous side, both, or in the phase lock material itself somehow?)

4) RNA precipitation: RT isopropanol followed by incubation at cold temps (~-20*C or lower).

• CHANGE: May omit this cold incubation and immediately spin down after adding RT isopropanol. A colleague said this would precipitate less junk.

5) cDNA synthesis kit: I prefer Bioline Sensifast ($ reasons), but have tested several kits.

• CHANGE: Add Qiagen Omniscript to the comparison.

Again, does anyone else have strong suggestions?

I don’t have the time, training or motivation to run a full design of experiments on this protocol, so I’ll be testing all (instead of each) of the reasonable and compatible changes at once and hoping the problem is eliminated.

Dear Brian

you seem rigorous in your approach. Im not sure what Im about to say unequivocally gets to the heart of the matter but nonetheless as above might constitute useful advice anyways:

1. I agree about gDNA contamination: I note from your first fig that the -RT control has a Ct approximating to 30-35 cycles; Ideally your -RT should be >/=40 cycles. Obviously in the absence of RT your only DNA for amplification must be genomic

2. In that regard whilst I think your basic approach to purifying RNA is excellent In my experience significant genomic contamination is most efficiently dealt with by DNASe I digestion: For effective genomic digestion, especially if your starting material is mg of tissue (as opposed to 1-2 million TC cells) then digestion is a pre requisite and furthermore digestion with a hyperactive engineered recombinant DNASe principle from Ambion works best of all and what is more results in NO visible RNA degradation (unlike other purified DNAses):

https://www.thermofisher.com/order/catalog/product/AM2238

3. What is more, if your primers are situated within exons; that is do not span exon-exon boundaries which would specifically target mRNA then this could undermine the efficiency of your PCR, which could in part explain the effect you are seeing with your pre diluted cDNA samples following PCR:

Have you performed a titration series depicted in your first figure with primers to a  high copy number   housekeeping gene ? Have you looked to see how the primers you mention perform in terms of standard RT PCR and in particular extra bands associated with genomic DNA. In general and for reasons of GLP I tend to optimise and select primer pairs for qPCR based on initial performance on agarose gels, i.e. RT PCR

4.  Looking at your problem in its broadest sense, inefficient RT; low copy number target and low PCR efficiency (due to side reactions associated with primer sequestration by gDNA) could potentially -at least in part - contribute to the manifest effects you are seeing

5. With reference to efficient RT, make sure your starting total RNA has a 260/280 of 1.8-2.2 and perhaps more crucially make sure your 260/230 ratio is . 1.0 and ideally > 1.5 indicating low phenol (in trizol) and /or GITC contamination (also from Trizol) both of which could contribute to poor RT efficiency

6. In terms of purification practices touched upon in your second answer/qualification of your original question) I am not a fan of precipitating RNA with isopropanol as cold temp with isopropanol culminate in excessive precipitation of genomic DNA and crucially salts including GOTC which reduces your 260/230 ratio and undermines the efficiency of your RT reaction. I routinely precipitate genomic DNA with room temp isopropanol but in the case of RNA I use 1/10 vol 3M sodium acetate pH 4-5 coupled with 3 x volumes of ethanol and incubate @ -80C for a minimum of 1 hour. Even @ such low temp ethanol precipitation is less prone to precipitating salts and crucially GITC in particular. Following precipitation a wash with ROOM TEMP  70% ethanol will also facilitate desalting (contrary to received wisdom, RNA is relatively stable as a pellet and thus room temp 70% ethanol can be used for washing and solubilises any remaining salt more effectively than cold ethanol)

7. Finally, in terms of efficient PCR - >85% is a pre requisite for accurate relative gene expression assayed by qPCR - keep your amplicons below 200bp (and id4eally ~100bp); make sure you select primers that span exon boundaries; and try and situate your PCR amplifications in the last 2KB of your mRNA as those 3' regions of mRNA replete in your cDNA population are always better represented

I have covered a lot of ground; I hope some of this will prove useful

As a minimum start by digesing your RNA with DNAse I; target exon-exon boundaries in terms of primer design; don't precipitate your RNA with cold isopropanol; screen primers initially on agarose gels to verify specificity and if you havn't already done so perform the studies you  show in your first figure with a ubiquitous high copy number house keeping gene to rule out copy number effects

Good luck !!      

Mr. Hall, thank you for your detailed reply.

To go point by point:

1-2: I’ll see about finding a way to include a DNAse step. The Qiagen cleanup with a DNAse digestion I've used here is not a step I usually include, but for reasons you’ll see in ‘3’ such a step may be required for one gene I’m interested in. it may be a good practice for me to adopt.

3: Most of my primers are designed such that they flank an intron (forward primer in one exon, reverse primer in the next exon), generally using a commercial program such as IDT’s real time primer design tool or Roche’s UPL tool. Unfortunately, a few of the introns these primers flank are fairly small (~150-300bp), and I suspect that reactions can amplify across the intron when there is contaminating gDNA present. Other primers, such as ‘BACT’ or Beta Actin seem rather specific for cDNA, which is the gene that was used for the original figure, and which was tested in the DNAse/purification test figure.

5: We quantify our RNA using a nanodrop. The 260/280 for our samples is general at or above 2.0. However, we sporadically have high 230nm absorbances, which I’ve always assumed might be due to phenolate or guanidine salt contamination. The absorbances were as below (based on a low N, I don’t record 230nm for all my samples because I’ve known it was sporadically bad for a while). I’m rather interested in solving this problem and may add an additional CHISOM wash to try and solve it.

230nm Absorbance after the Qiagen Cleanup:

Average: 1.5 | Maximum: 2.15 | Minimum: 0.15

230nM after Trizol Purification:

Average: 1.07 | Maximum: 1.37 | Minimum: 0.83

6: I’ll consider testing your ethanol precipitation protocol. I’m aware that ethanol is a less brute force method of precipitation, and I’ve used it for previous nucleic acid work. It certainly seems worth testing.

7: Noted. I can always design my primers for the 3’ region if there is a suitable intron. Especially if the programs I use recommend primers in those regions that have good TM, 3’ end stability, etc.

Addendum: Again, thanks for the help. We do check our primer’s specificity using a melt curve, which serves a similar purpose to agarose gel electrophoresis. That said, I could always do both until this problem is solved. As before ‘BACT’ is ‘Beta Actin’, which is a fairly highly expressed control gene (CT