The Bradford assay is simple, but can display considerable protein-to-protein variation, depending on the sequence, as it is sensitive to basic amino acids (Arg, Lys, etc.).

My mentors always encouraged me to do BCA protein assays, which are a bit more tedious, but in principle less sensitive to protein-protein differences, though Cys residues can have some influence.

Insulin is quite small and much smaller then BSA. Frankly, I'd be concerned about quantifying it with a Bradford assay based on BSA as a standard. I'd try to come up with some sort of positive control, testing a known quantity of insulin, before I'd consider the Bradford results reliable.

Thanks Peter for your advice. But I'm confused with the part "come up with some sort of positive control, testing a known quantity of insulin". May I know what actually does this means?

My goal would be to test some known concentration of insulin and see if you get reasonable results from the assay using your BSA standard, before doing assays to measure unknown amounts from cells/animals/what have you.

For example, purchase/obtain a known quantity of insulin, dissolve it in a known volume of buffer, and measure using your Bradford away, against your BSA standard. See if the result matches the known value.

I had done so and the results is almost half of the known value though. Is it possible to time a certain correction factor to the value that could correlate the number of amino acid in BSA and that of insulin?

I'm not sure one can derive a correction factor and I'm not sure it's safe to assume the error is a constant factor.

If you have known quantities of insulin then you may have to just prepare standards with that. Alternately, try another assay which should be less sensitive to these factors, at least in theory, like BCA.

The most reliable method to qunatify the amount of purified protein is to measure absorbance at 280 nm under denaturing condition (>3M GuHCl) and compare the result to the calculated molar absorption coefficient. The result of Bradford assay is strongly dependent on the amino acids composition, thus not adequate for studying e.g. Kcat or number of disulfide bonds per one protein molecule, etc. The disadvantage of the absorbance measurement is that high amount of protein is required to accurately measure the protein concentration.

Please refer to the paper 'Calculation of protein extinction coefficients from amino acid sequence data. Anal. chem. 182, 319-326 (1989)'.

Alas, if it was only that simple... Of the 20 common amino acids only three absorbe light at 280nm, Tryptophane, Tyrosine, and Phenylalanine. The absorbence of protein X will depend strongly on the relative content of these amino acids. Detrmination of protein in 280nm has its own biases, they are only different biases from those of Bradford or BCA.

Yoram

Is there any method to quantify protein by Bradford assay without using other standard protein such as BSA? The bias actually comes from the constraint that it is positively necessary to use a standard protein because the amino acid compositions between BSA and my protein would be different. The UV absorption method for protein quantification does not rely on using a standard protein, but only requires the extinction coefficient that can be accurately obatined by calculation.