It has been a routine practice to estimate a protein's amount by preparation of a standard calibration curve by BSA and with a spectrophotometer, followed by comparison with absorption of the sample protein. But I have been confused if this technique is to determine the concentration of protein (in molar) or the amount of protein (in milligram per litre). I have heard of both approaches.

 

When I check how the component Coomassie blue in Bradford reagent work, it is found that this dye binds to protein molecules proportionally (not dependent on protien length), so I would then think using this technique to deduce protein concentration would be a suitable one, but I couldn't understand why people use the same technique with weight determination (or weight determination only effective when calibration protein got a similar molecular weight with sample protein?). Hope for any help. Thanks.

I have been confused if this technique is to determine the concentration of protein (in molar) or the amount of protein (in milligram per litre). I have heard of both approaches.

 

This is equivalent to each other. Both are concentration determinations. You can calculate from the molarity to the weight, if you know the MW of the protein, of course. What you cannot do is calculate the MW of an unknown protein. For that you would have to weight in your protein (i.e. your calibration curve is then based on mg/ml) but then you cannot calculate the moles (again, if the composition of the protein is unknown).

This is equivalent to each other. Both are concentration determinations.

 

But if the dye binds protein proportionally, says, 1 dye to 1 protein; and says 20mg of BSA contains X mol of BSA; 20mg of my sample might contain Y mol of protein; so X mol dye bind to 20mg BSA, and Y mol of dye bind to my sample (20mg); though their weights are the same, but the color intensity (due to amount of bound dye) will be different. How could this be bypassed?

Incidentally it does not only scale with the amount of proteins but is also largely dependent on composition. The assay is not that accurate for complex protein mixtures (no existing assays are, btw.). Ideally you would always make your calibration curve with the protein or protein mix of interest, but this is rarely done. In the end you will always have to live with a certain margin of error.