I prefer Coomassie because you stain the actual gel BEFORE transfer.  Ponceau gives you an idea of how evenly your proteins transferred to nitrocellulose PVDF membranes, so if you have uneven transfer across your lanes, they will appear to be unevenly loaded when in fact that might not be true.  Coomassie can also be used after a transfer procedure because normally the amount of protein you load is far in excess of what will transfer, so you should still be able to tell if there is uneven loading.

If you have enough material, I would run a test gel with your lysates, stain with Coomassie and see once and for all if you loaded the same amount.  If you do, then the reason for the uneven loading controls could be:

1. Uneven transfer (unlikely).

2. Antibody not in excess (unlikely).

3. They are actually being regulated in these cells.

Has this happened in other cell lines you've used?  It might be worth making lysates from perhaps 5 plates of the same cells, load them up and do a b-actin/Ku70 blot under the same conditions.  I would also stain that gel with Coomassie.

Good luck!

Try quantifying proteins with BCA method for double comfirmation. And also, which proteins are you using as loading control?

Hi there,

What you are actually looking for is a loading control for normalizing your quantitative data, this page should help you:

http://www.abcam.com/primary-antibodies/loading-control-guide

Good question from Madhur - so-called house-keeping proteins used as loading control proteins can sometimes actually be up- or down-regulated depending in your conditions. 

Have you tried staining your gel with Coomassie Blue?  This would at least give you some idea on whether you ARE actually loading the same quantity of protein into the gel.

I am using a loading control. I am currently using B-actin but have also tried Ku70 ... neither of which have equal amounts according to my blots. I had considered this but feel like it is unlikely that they are both altered in my cells. I have stained with ponceau as well, is this similar to the Coomassie blue? I had been told they were more or less the same.

Techniques wise your protocol seems good to me. Do you expect neurospheres to have too much lipids. They are known to over estimate protein concentration and incase some experimental wells have more lipids, it might lead to uneven loading. This is the case with adipose tissue and adipocytes that we use and have to make sure to add sufficient amount of sds during lysis. Which lysis reagent do you use?

I prefer Coomassie because you stain the actual gel BEFORE transfer.  Ponceau gives you an idea of how evenly your proteins transferred to nitrocellulose PVDF membranes, so if you have uneven transfer across your lanes, they will appear to be unevenly loaded when in fact that might not be true.  Coomassie can also be used after a transfer procedure because normally the amount of protein you load is far in excess of what will transfer, so you should still be able to tell if there is uneven loading.

If you have enough material, I would run a test gel with your lysates, stain with Coomassie and see once and for all if you loaded the same amount.  If you do, then the reason for the uneven loading controls could be:

1. Uneven transfer (unlikely).

2. Antibody not in excess (unlikely).

3. They are actually being regulated in these cells.

Has this happened in other cell lines you've used?  It might be worth making lysates from perhaps 5 plates of the same cells, load them up and do a b-actin/Ku70 blot under the same conditions.  I would also stain that gel with Coomassie.

Good luck!

You may already be doing this, but I always recommend counting your cells before lysis.  Then adjust the volume of cells you spin down so that the same number of cells is in every tube.  If you do this, the concentration of protein you get should be very similar across samples and you should be loading almost the same volume of lysate in each lane.  If this is true but you still get big differences in the loading control, then  perhaps the various conditions are affecting the level of your housekeeping gene/cell or even total protein/cell.  If it's possible, you might divide a single cell suspension into several aliquots, prepare lysates, and load each one on a gel.  This would give you an idea if the variability is just a technical issue.

to get evenly diluted samples for sds-page, your protocol should look like than sample x mcl, diluent y mcl, and 10 mcl 4xsample buffer, where x+y equals 30 mcl, and x contains 40 mcg protein, and less/equals 30 mcl in volume. and use protease inhibitors in your diluent/lysis buffer. what is your lysis buffer? it looks like you use PBS only.

Use house keeping genes like GAPDH or Actin. Then normalize your protein test sample with control, that you can do by densitometry.

Coomassie stain is usually not compatible with Western blot as it os irreversible. There are modified CB stain as well as other staining method- I don't have materials in my hand now but will come back. I believe it's rather protein quantitation problem. Once I had a contamination from my culture (BSA in culture media and proteins from ECM matrices) and failed to have equal loading- ended up with throwing away those samples. 

Have you tried loading with tubulin to see if the results are due to loading error vs. actual results. Tubulin is not affected by other conditions and will therefore be a good indicator of uneven loading. You can normalized based on that.

Which protein you are using as a loading control. I have used tubulin and its expression is uniform in cancer cell lines. Please also be very careful with the quantification of cell lysate. Are you using same lysis buffer to lyse neurospheres that you have used for monolayer lysis, that can also be another reason.

Hi, 

basically your method is fine, I would not recommend using Coomassie since you do want to have equal amount of protein on you membrane.

if your bradfored curve is OK, I will use Actin or Tubulin as a loading control. if you have Odyssey you can even try to stain your membrane with both you protein antibody and the reference (Actin/Tubulin) and get quantitative ratio between them in each lane, thus normalize your results.

I don't really see what the problem is. Every protein quantification method has its drawbacks and incompatibilities. But, even though the method you chose is the perfect one, pipetting errors can happen, due to the operator and to the micropipette. The point is that in my experience it is normal not to get even protein loading in WB, but that is why we use loading controls (house-keeping proteins and total protein stainings).

Unless you have checked a certain house-keeping protein in your system actually does not change in any given circumstance, I would recommend to use a total protein stain. I use Coomassie for staining membranes after I finished all the blots and it works fine.

If the cells are growing as neurospheres, there could be a lot of cell death, or extracellular matrix production, or other odd events which are skewing your protein assay.  Perhaps try loading equal volumes of lysate and see how that works out (or count the cells first and then load equal volumes).  You could also try loading 5, 10, 20 and 40ug of the same lysate, and check that the bands roughly double in intensity to tell you whether the relationship between your staining and what you think you're loading is roughly linear.

I will try counting cells before I lyse them and I may also change my lysis buffer from NP-40 to RIPA juat to see if this improves things. I will also try staining with Ponceau. Thank you for your help.

How uneven are your bands? In my experience, variance up to 25% should be expected. 

Variations are quite substantial, often more than 50% difference.