Western Blotting vs ELISA

Giovanni Maga maga at vetbio.unizh.ch
Tue May 16 09:59:24 EST 1995

In article <3p047d$l3l at nnrp.ucs.ubc.ca>, hoekstra at unixg.ubc.ca (Sucking
Gippy) wrote:

> Currently, I'm running a number of SDS-PAGE's, tranferring to 
> nitrocellulose, and immunoblotting for protein of interest (stress 
> proteins). 
> I understand that western blotting gives more of a representation of 
> proteins presence/non-presence and was wondering if investigators 
> ROUTINELY run ELISA's to quantify (i.e. get a numerical value for 
> comparison) the protein of interest, as compared to a control sample? 
> In my mind, there is no doubt that the protein is present...are ELISA's 
> necessary then in addition? or is something like using a densitometer to 
> compare present bands sufficient?
> Thanks.
> PS: If anyone is personally researching stress proteins (ie HSP 70 etc.) 
> I would appreciate your reply..

Western blot is of course more representative of your protein since you can
clearly identify the band, thus having direct confirmation of the MW. It
also excludes the possibility of false positives. I would say that for a
clear-cut demonstration WB is needed in addition to ELISA and not
vice-versus. ELISA test is usually run as a routine screening technique for
a large number of samples (for example when you are producing MAbs and want
to identify the right clone) which is problematic with WB. Quantification
of proteins from a WB is more tricky. Densitometric scanning of WB bands
(for example obtained with ECL system which gives you an autoradiogram
easier to be scanned) is sensitive, given that you compare control lanes
and positive bands detected with the *same* Ab. You can make a titration,
using different known amounts of pure Ag and comparing the intensities with
the one obtained in your unknown sample (better if you do a serial dilution
of the unkown sample in order to get better estimate and to check the
linearity of the densitometer response). The thing you must be careful
about is to use *exactly* the same conditions for WB with the control and
the unknown sample (Ab concentration, blocking, washes, incubation time,
membrane, secondary Ab, etc.) in order to minimize fluctuations due to
experimental conditions. I would suggest to run a first experiment to get a
rough estimate of the best range of concentrations and then make the best
one to get the value you need (better if you run a couple of independent
Regards, G.Maga.
maga at vetbio.unizh.ch

More information about the Proteins mailing list