Northern Normalization

Bernard Murray bernard at
Sat Aug 24 23:20:39 EST 1996

In article <tyou.319.321E955A at>, 
tyou at says...
>In article <mantei-2208960800050001 at> 
mantei at neuro.biol.ethz.c
>h (Ned Mantei) writes:
>>In article <321B12ED.304E at>, Heinz-Juergen Schaefers
>><mfm434 at> wrote:
>>> the best internal standard for RNA loading is 18s or 28s rRNA.


>The other way:
>1. Duplicate lanes with same amount of RNAs (polyA+ or 
>total)2. Cut the filters in two
>3. Probe each filter with two different probe, one of which is your specific 
>probe and the other is well known control gene.
>I know some people used ribosomal protein gene probe in plants. Soybean actin 
>gene comes to mind also.
>4. Using densitometer or computer analysis, you can normalize your signal 
>relative to the other control intensity.

This method will compensate for sample to sample differences in RNA
purity and integrity (you are trying to load the same mass of RNA)
but will not compensate for either variation in sample loading or
unevenness in the transfer.  To cope with this the accepted way is to
use the equivalent of an "extracted standard" and this is accomplished
by probing the blot for your signal of interest and then stripping and
reprobing with the standard.  If you are lucky and are sure that your
probes give you a clean signal you can apply both at once (assuming
the sample and standard have sufficiently different electrophoretic
mobility) but smearing of the bands can make this difficult to interpret.

I agree that RNase protection can be one of the best methods for
quantification especially if the RNA can be hybridised to the sample
and control templates simultaneously and then the products separated
on the gel.  I like GAPDH or cyclophilin as controls in this method
(actin or 18S also work acceptably) but none give a rock-steady baseline
especially when analysing RNA from cultured cells (throw in some
cycloheximide and watch the level of 18S shoot up!).
	I try and restrict Northerns to checking the size of an RNA
and normalise by 18S rRNA (total RNA) or GAPDH or a ribosomal protein
mRNA (polyA selected).  I also feel that methylene blue staining of the
filter and then densitometric analysis is also an acceptable approach.
	Just my opinions...

Bernard Murray, Ph.D.
bernard at  (National Cancer Institute, NIH, Bethesda MD, USA)

More information about the Methods mailing list