How to get the correct standard BSA curve for my protein estimation?

I have purified my protein and want to estimate the amount. The plot I am getting is half of the value my senior is getting. The difference is I am using a kit (Genei TM) for Bradford's assay. But he prepared his own reagent. There might be a difference in reagents, which I can understand, but the amount of protein should be the same and I am only getting half the amount of protein as earlier estimated by my senior. Please suggest what I should do. The dilution I am making is not wrong and the expiry date of Bradford's reagent from the kit is Dec 2013. So how should I proceed now? The slope value he (senior) got is 0.0196 and the R value is 0.99 from his own reagent. My slope is 0.009 and my R value is 0.98


0 / 0  ·  36 Answers  ·  6112 Views

Popular Answers

All Answers (36)

  • Aleardo Morelli · University of Minnesota Twin Cities
    Why don't you try to use your senior's reagent or have him use the kit and see what happens. There might be a difference between the home made reagent and the commercial one, and if these two react differently the results will be different.
  • Alvaro Olivera-Nappa · University of Chile
    Hi, Prem.
    The Bradford assay is very useful and fast, but its chemistry is not truly irreversible since it is an equilibrium reaction. It depends on the reaction of charged groups on the protein surface, mainly arginine, but also histidine, lysine, tyrosine and tryptophan residues. Therefore, it depends on the reactant concentration, temperature, and pH of the sample, as well as any other condition that could affect the reaction, including the specific buffer and salts included in the reagent. If the reagent concentration is different between assays, the results could be different. Besides, the equilibrium seems to be fast, but for some proteins the absorbance changes if your incubation times are longer.
    Besides, this assay is non-linear due to the reversibility of the equilibrium reaction. To get proper results and linear standard curves, you should use one of the linearization methods that takes into account the ratio between absorbances at 590 nm and 450 nm. Check if you are interested in this and also on an explanation of the equilibrium reaction.
    You should also consider that the Bradford assay is not affected by cations or sugars, but it could be strongly affected by detergents and alkaline solutions. This could be reduced when considering the ratio of absorbances instead of using only the absorbance at 590 nm. This consideration also would reduce the differences caused by reagent concentration effect on the equilibrium.
    To complicate more the problem, BSA is known for being one of the exceptional proteins when performing the Bradford assay. It is cheap and readily available, but it is usual for it to give around half the reactivity to Bradford reagent than other proteins (for example, IgG). Moreover, depending on the source and purity of your BSA sample, the results could be different. This makes BSA a poor standard for this assay. You could use IgG, ovalbumin, lysozime or catalase instead.
    In conclusión, I would recommend you to use another standard instead of BSA and to carefully check the concentration of the reagents. If these concentrations are not the same, the results will not be the same, but I am pretty sure that if you use the 590/450 absorbance ratio instead of the crude absorbance at 590 nm, the results will be more similar.
    Hope this helps. Good luck with your experiments!
  • Sukanta Bhattacharya · University of Cincinnati
    I would suggest you prepare a fresh stock solution of BSA and go for sequential dilution from the stock. Proceed with the protocol of protein estimation(whatever you feel comfortable) both for your standard as well as samples to be estimated. Check the OD of the sample against that of your standard plot and you will get the correct result.

    For accuracy, it is always suggested that you prepare a standard plot every time you estimate the protein. I hope this helps! Good Luck!!
  • Roger Rowlett · Colgate University
    If you want analytically accurate measurements for protein concentration I would not use the Bradford assay. It is subject to many interferences, and is highly variable depending on hydrophobic content of the protein assayed. Additionally, BSA is a poor choice for the Bradford assay standard because it has an anomalously high hydrophobic dye binding capacity. I use lysozyme or some other protein with an average hydrophobic dye-binding capacity instead if I must use this assay. If you are using different dye reagent lots, this is likely the cause of the differences you are seeing. Also, it is an easy mistake to use the wrong Coomasssie blue dye (there are many dyes that go by that general name) when preparing your own Bradford reagent.

    The least idiosyncratic protein assays are the A205 (Scopes) UV method, or the microbiuret assay (in the absence of interferents such as thiols).
  • M. Rajamohamed Kalanjiam · Alagappa University
    Check the accuracy of the spectrophotometer used in your lab
  • Thierry Michon · French National Institute for Agricultural Research
    Just a remark.

    BSA is certainly not the good standard as it over-responds to Bradford.
    Gamma globulin is far better.
    let's take a 10 mg/mL standard solution and you estimate its concentration by three different assays (Biuret, Lowry and Bradford)
    here are the results :
    BSA (Biuret) : 9.7/ (Lowry) : 8.4; (Bradford) : 21.1 !!!
    Gammaglobulin : (Biuret) : 9.4; (Lowry) : 11.8; (Bradford) : 8
  • Jacob McMillan · The University of Memphis
    I have had major problems getting accurate concentrations in the past with the Bradford Assay. I have since switched over to a BCA assay kit from Thermo-Fisher (I'm sure other companies also have kits), which is more compatible with my protein system and have gotten much better results that are consistent with other quantitative methods.
  • Jorge Ventureira · North Carolina State University
    Maybe you are making an error in calculations. slope from you is half from your senior and your result is the half. Is your sample abs 1/4 compared with that of your senior?
  • Gnanasekaran Gopalsamy · Chonnam National University
    Protein estimation is not accurate by using any method (bradford or lowry) available in web-site. Protein estimation is depend on aminoacids. Some method recognize particular aminoacid and estimate the concentration. You can measure your protein concentration at 280 nm and also estimate your protein conc. on SDS-PAGE by comparing protein ladder. You can use all methods (bradford, lowry, 280 & SDS-PAGE) and estimate your protein concentration. Good luck.
  • Alok Nahata · Dr. Harisingh Gour University
    I will suggest you to make your own reagent using the same procedure as followed by your colleague. You can use your own chemicals or borrow the same from him. This will allow you to have a comparison with the colleague also and will also help to find out the problem with the kit reagent, if any. So there will be now three reagents with you, 1. Prepared by you 2. Borrowed from your colleague 3. Your kit.
    Once you get the readings from all of them, you can arrive at a fruitful decision. Only thing will be a repeatation of the experiment with all available options.
    Pls inform about ur results :)
    Gud Luck
  • Adrian Velazquez-Campoy · University of Zaragoza
    If you are using different reagents/dilution, it is no surprise you get a different standard curve. This is not a problem at all. But, the important point is if, when estimating the concentration of an unknown sample by interpolating into your standard curve, you get the same final (mg/mL, not OD) value as your colleague.
  • Some good answers above, but I am a little confused by the statement "The difference is I am using a kit (Genei TM) for Bradford's assay. But he prepared his own reagent." If that means you both used the Bradford technique, but you used a kit reagent, and he used a lab-prepared reagent, this could be a major discrepancy. Buffer components , especially detergent content, can also be a major source of discrepancy, as can source and purity of BSA. I have employed many protein estimation methods, but finally stuck with Pierce BCA kit due to low protein variance (stated as 15%), low buffer variance, and low interference by detergent. For a more linear curve, truncate the standard curve between 0.1 - 1.2 AU per ug/ml protein. The BCA instructions have some good information on the above:
  • Santhosh Kumar · Loyola College
    Ya I too agree with Adrian Velazquez-Campoy. The slope or R value > 0.90 is not a problem at all. To check whether your equation is correct or not, you can use the OD of your Known standard (say 2 mg/ml) and use it in the equation. if you get any thing near 2 (98%) you have done correctly. Provided you should have used the same reagent for your unknown samples also.
  • Lakshmi Sunkara · Oklahoma State University - Stillwater
    I have been using Brad Ford Assay with Bio-Rad kit and I did not get any problem. Have you diluted your original protein similar to your senior did?
  • Melvin Klegerman · University of Texas Medical School
    Bradford is a protein dye-binding method and can vary greatly from one protein to another. One of the functions of albumin is to bind small molecules and usually binds more dye than other proteins, giving higher ODs for the standards. Hence, your protein will assay lower. All protein assay methods give variable results, depending on the protein. It is important to match the standard as closely as possible to the sample to be assayed, but requires some understanding of the assay principle. I published a paper on the subject: Klegerman, M.E., Hamilton, A.J., Huang, S.L., Tiukinhoy, S.D., Khan, A.A., MacDonald, R.C., and McPherson, D.D.: A Quantitative Immunoblot Assay for Assessment of Liposomal Antibody Conjugation Efficiency. Anal. Biochem. 300:46-52 (2002).
  • Inna Gorshkova · National Institutes of Health
    The first thing I would recommend is to reproduce your colleague's result using his chemicals, his pipettes and his conditions. Then we can decide what to do next.
  • Engelbert Buxbaum · Ross University, School of Medicine
    There are several parts to the answer:

    1) Most fundamentally, there is no such thing as a correct protein concentration. All commonly used methods make relative measurements, i.e., you result is: This assay is as blue as would have been obtained with, say, 1 mg/ml of your standard protein. However, different proteins with different amino acid composition give different color yields. Depending on which combination of assay method and standard protein you use, protein concentration for the same sample can differ by a factor of 20 (been there, done that: Lowry vs Bradford and BSA vs IgG as standard and Hsc70 as sample).

    2) Apparently, you can now get IR-LED spectrometers that read the Amide I band to get an absolute (independent of aa composition) estimate of peptide bond concentration, but I have not used them and cannot comment on interfering factors.

    3) If you know the amino acid composition of your protein, you can calculate its absorption coefficient at 280 nm. There are two methods to do this, Perkins 1986 and Gill & van Hippel 1989. The results in my hands agree to within a few %. However, this obviously does not work for protein mixtures.

    4) Do not use BSA as standard protein for Bradford assays, BSA has binding sites for hydrophobic substances and gives more intensive staining with Bradford reagent than other proteins. IgG behaves more typical. The Bradford assay has a fairly narrow linear range, parabolic rather than linear regression of the standard curve may be more appropriate. As with all photometric assays, make sure that you have no precipitate, gas-bubbles or scratched cuvettes that would scatter light.

    4) In your case, even if both reagents gave different color yield, this should cancel because you always use standards with Bradford's method. I.e., if your reagent gave, say, a lower color yield than that of your colleague, it would do so for both the standard and the sample, and the effect would cancel out. I therefore suspect that either you are using different standard proteins, that either one of you made an error preparing your standard solutions, that you are using different wavelengths (optimum is 590 nm with an isosbestic point at 535 nm), or that there is a calculation mistake somewhere.

    I'd suggest that you do the assay again, side-by-side with both your reagents and your standard solution (that is, a total of 4 assays).
  • Juan Pablo Fernández-Trujillo · Universidad Politécnica de Cartagena
    Please be careful If you have a very low concentration of protein in your samples for any reason. We adapted the Bradford method with the Biorad kit for apple protein. Some cultivars had the standarc concentration but others (crabapple) very low concentration. Using 500 uL extract I also got good results. If this was your problem after following the good advices mentioned above, you can check: Fernández-Trujillo, J.P., Nock, J.F., Kupferman, E.M., Brown, S.K., Watkins, C.B. 2003. Peroxidase activity and superficial scald development in apple fruit. J. Agric. Food Chem. 51: 7182-7186.
    The other choice if the problem would persists is trying to concentrate the protein or to desalt peptide or protein solution.
  • Salvatrice Rigogliuso · Università degli studi di Palermo
    How to get the correct standard BSA curve for my protein estimation?
    In my experience I know that each time that you make or you buy a new reagent is need to prepare a new standard curve. Each reagent infact can be different in the ability of estimation. Moreover, because for example the Bradford reagent use a colorimetric reaction, often it can be oxidate and give a different reading of the same sample. I suggest you to prepare a new standard curve using known concentration of BSA, for example 2mg/ml, and from this prepare a serial diluition. (about six different concentrations). You can used this standard curve until the same reagent will be used. When you would buy a new reagent, you must to prepare again a new standard curve.
  • Essam Al-Jumaily · University of Baghdad
    please read carefully the Bradford method which pub. in (1976), because there are reagent you must parper first standard protein (BSA) between ( 0-100 ug/ml). and parper the cooamassie blue G-250 reagent ,also each time you used this reagent you must parper standard curve. and I think your problem with a kit .
  • Freddy Dardenne · University of Antwerp
    This is a long ongoing discussion because of various reasons. I list a few. Everybody who is using Bradford has long forgotten how it really works (when was it first described?). Actually it binds aromatic amino acids, mainly tryptophane. So proteins with low numbers of these amino acids will react poorly in a Bradford assays. A good example is RNaseT1 which will practically show no reaction at all. Statistically this problem gets smaller as your protein gets bigger or if you're measuring mixtures of proteins. This also means that for purified protein the best standard is the protein itself or a very similar one. This holds for every assay with speciific binding or reaction characteristics. In any case if you compare results from different colorometric assays using different standards you will practically never get the same results and differences can be very big. Also look at temperature in the lab or incubator, matrix effects, buffers used, .... By the way, You can use Bradford calibration lines for a longer time if you control your assay strictly, including reaction time.
  • Jeffrey Boles · Tennessee Technological University
    I have one additional insight in addition to the above. Did you make your own reagent? We have observed that if you use Ethanol, you get about half the absorbance as when you make the dye solution with Methanol, thus, we always use Methanol. We've assayed protein with commercially available dye and get the exact same response. Secondly, when you make your standards, you need to have an accurate technique for measuring the actual concentration (such as a good extinction coefficient, etc.). You cannot simply weigh it out and assume it is correct. We've discovered in our Proteomics research that most of the time, BSA works well, but sometimes it's best to use the target protein in the preparation of standards.
  • Indesh Attri · CSIR - Institute of Himalayan Bioresource Technology
    I am also working on proteins and i have doubt in my mind that the protein absorbance at 595 nm using Bradford reagent the OD values of BSA standard curve above 0.600 are not considered is that true???If yes then also explain this why?
  • Freddy Dardenne · University of Antwerp
    To Indesh. Cutting of at an OD above 0.6 (or any other for that fact) is in principle not necessary. Increasing the protein concentration in the assay will eventually lead to saturation, showing a curve like a Michaelis Menten Kinetic. You can use a standard curve outside of the linear range, but obviously in the error of the calculation will increase as the slope of the curve becomes smaller. Next to that also yur spec will have a linear range, above that the photomultiplier is hit by so few photons that the multiplication becomes very erroneous. Keep in mind these instruments actually measure transmission, not optical density..
  • Rakesh Nair · Catholic University of Louvain
    I would suggest you to use the Bicinchoninic Acid Protein Assay Kit from Sigma Aldrich for protein assays, its pretty reliable and no need to make reagents on your own. Also it would be better to always run your standards along with a blank for all spectrophotometric assay. Both you and your colleague can perform the assay together separately at the same time using the same set of reagents, standards and samples (dilutions) and check whether it was not the case of manual error. That could give you a much better solution to your problem.
  • Indesh Attri · CSIR - Institute of Himalayan Bioresource Technology
    I am using BSA standard with conc. ranging from 10 to 50 micro gram 40 microgram OD value is coming 0.6 so i should stick to 40 microgram range for the calculation of my sample ?? Please suggest.
  • Freddy Dardenne · University of Antwerp
    Like I said, this all depends on the shape of your curve. Did you test where you leave linearity? And even beyond linearity, what accuracy do you need/want ?
  • Ashok Dubey · Netaji Subhas Institute of Technology
    BioRad protein assay kit has worked pretty well in our hands. We follow exactly the same protocol as supplied by the manufacturer. However, measurments must be done carefully and precisely. Given that, one should be more confident about one's own observations, which may be in disagreement with results reported by others for various reasons.
  • Nirpendra Singh · Dr. B.R. Ambedkar Center for Biomedical Research
    Dear Prem,
    I feel there must be some error in the plotting of your value on the excel.
    Just let me know few point
    1) Did you and your senior have included the zero point in your graph,if you allow the line to pass from zero and your senior wont allow the line to start from zero this kind of error might take place, because in this case the slope will change and but the R value would be 0.98 or 0.99.
    2) Both of you might have used the same standard or did you prepare your own standard and your senior might have used the other standard,in that case the dilution faction might be an issue for standard.
  • Aleardo Morelli · University of Minnesota Twin Cities
    Hi Prem,

    As Freddie pointed out, different proteins react differently with bradford reagent, and best results are obtained by using a protein similar to yours to make your standard curve. If your protein is not similar to your standard, results might be completely useless even if you get your curve to work like your senior's. And I do not know if you have done so yet, but trying using your senior's reagent, or get your senior to use the commercial kit might not harm, and will answer the question: the discrepancy in the standard curve is due to an operator or a reagent issue?
    However, if your goal is measuring the exact protein concentration I would suggest that you send the sample out for aminoacid analysis which is far more reliable than any colorimetric assay.
  • Jiang Jiang · Jiangnan University
    Yes, I agree with what Freddy Dardenne said. for example, if you want to test the content of soy protein solution, its protein contents will be determined by Bradford method using a Kjeldahl protein concentration-calibrated SPI as standard.
  • Essam Kotb · Zagazig University
    I usually use Bovine Serum Albumin standard with conc. ranging from 10 to 100 ug and give my straight line without any difficulty. be sure that you dissolve BSA well and use NaOH . try folin reagent and compare.
  • Juergen Wiegel · University of Georgia
    I am very surprised that BSA is still promoted that much,since it has been shown more than 30 years ago, that it gives for most enzymes too high values. (up to twofold high) You can use it but if you want indeed accurate values you need to establish a correction factor for your enzyme. The best way is to calculate from the sequence your nitrogen content and do a Kjeldahl . Otherwise use different protein standard to get a correction factor.
  • Catherine Tétard-Jones · Newcastle University
    and for whole proteome analyses - what does everyone think is best to ensure a fixed amount per sample is analysed? I noticed some studies use a fixed amount of tissue, and don't quantify how much protein it contains - would that be a better approach to quantify protein abundance (via 2DE)?

Question Followers (45) See all