I'm doing a western blot with tumor samples that have necrotic tissue inside. I used Pierce BCA protein assay kit to measure protein concentration. But some proteins are not consistent-- with the same amount of loading, actin bands were not always the same, some samples even varied from time to time. Should I use a different control? Any suggestions?
Don't get me started on normalization of blots but glad you brought this up! There is a tendency for blind usege of actin as a housekeeper in Westerns that came from the molecular biologists. Its totally appropriate for qPCR but not always for Westerns as you have to be absolutely sure that actin (or any housekeeper) is not changed in any of your samples by your treatments or,in this case, necrosis.Therefore in necrotic tissue vs normal it is very likely that it will be changed as will a lot of other proteins therefore no 'housekeeper' protein is appropriate. This is the same as looking at 2 different tissues and wanting to compare the percentage of the same protein in each. What you want is the percentage of specific protein compared to the weight of total tissue protein in your sample because one doesn't know if so-called housekeepers are the same in both tissues. So you have done exactly the right thing and what you are seeing is inherent variation in your tissue samples. Variations between Western runs with the same sample also could be due to other factors.....some proteins fall apart or aggregate on storage in SDS buffer so may change with time. Likewise, Westerns are inherently variable in transfer efficiency and sensitivity from day to day so that may be part of the issue. Stick to your protocols rigidly as even the time spent soaking the transfer membranes in solvent and transfer buffer can have an effect. As one of your replies suggested, staining in Ponceau red before probing will give confidence that the tracks are all loaded the same
Can you attach your blot? We may help you better after looking at it.
Don't get me started on normalization of blots but glad you brought this up! There is a tendency for blind usege of actin as a housekeeper in Westerns that came from the molecular biologists. Its totally appropriate for qPCR but not always for Westerns as you have to be absolutely sure that actin (or any housekeeper) is not changed in any of your samples by your treatments or,in this case, necrosis.Therefore in necrotic tissue vs normal it is very likely that it will be changed as will a lot of other proteins therefore no 'housekeeper' protein is appropriate. This is the same as looking at 2 different tissues and wanting to compare the percentage of the same protein in each. What you want is the percentage of specific protein compared to the weight of total tissue protein in your sample because one doesn't know if so-called housekeepers are the same in both tissues. So you have done exactly the right thing and what you are seeing is inherent variation in your tissue samples. Variations between Western runs with the same sample also could be due to other factors.....some proteins fall apart or aggregate on storage in SDS buffer so may change with time. Likewise, Westerns are inherently variable in transfer efficiency and sensitivity from day to day so that may be part of the issue. Stick to your protocols rigidly as even the time spent soaking the transfer membranes in solvent and transfer buffer can have an effect. As one of your replies suggested, staining in Ponceau red before probing will give confidence that the tracks are all loaded the same
Sensitive total protein stain (example Lavapurple) so you can accurately quantify the amount of protein in the whole lane.
In addition to Ponceau S, you could also try Coomassie staining the membrane once you have finished all the immunoblots, or you could run the samples again and use Silver Staining to stain the gel. Coomassie is more sensitive than Ponceau S, while Silver Staining is much more sensitive than either Ponceau or Coomassie stains (and therefore you can load very little sample on the gel and still get a strong signal). Each of these methods would give you a good indication of the protein content of each sample, which should give you a general sense of whether the samples are loaded evenly. But as John Wardale says above, using a specific protein to compare between different samples is not always a good approach, especially when comparing between different tissues.
Why do not you try ELISA? Today, so many proteins are being detected by ELISA. The western blotting seems to be replacing with?...
Hi Xi, you can try prerun your samples first using small amount of samples with any loading control you comfortable - actin, tubulin, lamin etc. You don't need to do full long run as Abs for those protein highly sensitive and proteins are highly expressed. Then normalize your load for main run. It will take more time but quality is much better.
Great question and good answers. As long as you don't try to find "the best" way you are doing well, I believe there is no best way, but the more controls you have, the higher the chance that you will not fool yourself.
I agree that staining the nitrocellulose with Ponceau red before blocking and incubation with primary antibody is always a good idea (make sure you keep a good quality digital image of the stain for later reference). Measuring total protein concentration and equalising these from the start is also essential, and for good western blots you should load equal volumes (with equal protein concentrations), rather than different volumes to compensate for differences in concentrations.
But on top of all this, I would still probe with a few house-keeping gene products that give good signals, so actin can be one. Whatever the variation in your sample physiology, there probably is no perfect house-keeping gene, but using several of them gives you different types of ratiometric outputs, all of which can be useful for further analysis and interpretation. So the more references you have, the better can you describe the behaviour of your protein of interest. May be your protein of interest is more abundant relative to actin, and less abudant relative to the ER chaperone BiP. Having both datasets simply gives you more information, and better composite figures for publication.
Actin is structural microfilament and usually localized in cytoplasmic part. When cell under go to necrosis it's eventually releasing all cytoplasmic contains so on actin perhaps be released too.
So, i suggest to use another protein such as vimentin that intact at cell membrane(i assumed that) or detect cell at early stage of necrosis if that possible
You can Normalize against total protein load, Much stable than single house keeping proteins.
- Badges
- Science method