Expect the Unexpected - RNase A may still catch you by surprise
by Federico Donà and Jon Houseley Labtimes 01/2015
Researchers thought they knew most everything about RNase A: the RNA degrading enzyme was purified in the 1940s already, the structure was solved in the 1960s and the catalytic reaction mechanism can be found in detail in every biochemistry text book. But RNase A may still catch you by surprise.
Some laboratory reagents are so familiar that we take them for granted and yet these molecular workhorses can occasionally surprise us, yielding unexpected and potentially misleading results in standard procedures.
We were not expecting RNase A, an exceptionally well characterised enzyme, to do anything other than degrade the RNA in our samples but we actually observed RNase A removing substantial quantities of DNA under common experimental conditions. This activity confounds the results of basic RNase susceptibility assays and has the potential to introduce sequence bias during DNA purification that would introduce artefacts in genome-wide analyses.
While searching for transcripts from mouse major satellite repeat sequences by Northern blot, we observed strong signals in RNA samples derived from cultured fibroblasts. To ensure that these signals did not represent contaminating genomic DNA, we performed a routine digestion with RNase A and the signals disappeared as expected. Other results, however, were puzzling as the major satellite signals were completely removed by DNase I but not by some other ribonucleases.
A major satellite PCR product is efficiently removed from an RNA sample by RNase A treatment and phenol/chloroform purification. RNase A from three different manufacturers was tested.
This unexpected result led us to suspect that the signals stemmed from genomic DNA and that RNase A may possess an innate activity against DNA. Confirming this suspicion, we found that RNase A readily removed the signal of a major satellite PCR product spiked into the assay (see Figure). This ability was not unique to a particular batch of RNase A and multiple other brands, including certified DNase-free RNase A, displayed the same activity.
Anecdotal reports do suggest that RNase A can degrade DNA but such reports receive little credence as the well-understood catalytic mechanism of RNase A absolutely precludes DNA cleavage. It was shown long ago, however, that RNase A can bind DNA and our experiments reveal that RNase A holds onto bound DNA with such ferocious tenacity that DNA-RNase A complexes are not dissolved by phenol:chloroform. When phenol:chloroform extraction is used to purify DNA samples treated with RNase A, these complexes partition efficiently to the organic phase and the DNA is lost.
Loss of DNA is not a peculiarity of major satellite sequences as a DNA molecular weight marker was also efficiently removed by RNase A. However, we predict that there will be a sequence bias as RNase A binds to single stranded regions of DNA as it ‘breathes’ (a process in which regions of the duplex transiently open), which occurs at a sequence-dependent rate. The quantities of DNA removed are also not small; 1µg RNase A completely removes at least 50ng DNA from a sample, so a few µl’s of commercially available RNase A (usually 10-20mg/ml) added to a DNA sample could remove micrograms of DNA.
There is no easy solution to this problem as proteinase K digestion was insufficient to completely release the bound DNA. The effect can be partially suppressed by increasing the salt concentration but DNA removal still occurs. Column-based purification methods should avoid DNA loss but may suffer the opposite problem of co-purifying the RNase A leading to DNA migration defects, which we also observed.
We recommend that RNase A treatment is avoided entirely for critical DNA purification applications, particularly on low abundance samples. RNase T1 did not show the same effect in our hands and should, therefore, be usable for RNA removal post-purification.
Of course, RNase A finds many applications outside DNA purification; we suggest that care is taken when interpreting any assay involving RNase A, to ensure that the results cannot be explained by strong binding of RNase A to DNA or chromatin.
Federico Donà and Jon Houseley
(The Babraham Institute, UK)
Last Changed: 04.02.2015