We found in our laboratory that the best way to measure protein concentration (regarding sensitivity and accuracy) is to use spectrophotometric method. We use a cuvette that needs 70 uL, but since the method is not destructive to the sample you can recover it. (There are even some cuvettes that need less.) The important thing is that when you measure small volume you still keep a 1 cm pathlength. If you measure in a microtiter plate the pathlength gets very small and decreases the sensitivity since absorbance is proportional to concentration and pathlength (Lambert-Beer law).

We measure absorbance at 228.5 nm and 234.5 nm. Protein concentration is determined according to formula: γ (protein) = (A228.5nm - A234.5nm)×0.3175 [mg/mL]

It is much more sensitive than measuring at 280 nm which is expected if you consider the absorption spectra for proteins.

Reference: Ehresmann, B., Imbault, P., Weil, J.H., 1973. Spectrophotometric determination of protein concentration in cell extracts containing tRNA’s and rRNA’s. Anal. Biochem. 54, 454–463.

In simple words, you got just ~2 micrograms of your protein. Any destructive method, requiring at least half that quantity (e.g. Bradford, Lowry and so on) is a desperate choice, i.e. if you must mandatorily know your protein concentration you have to sacrifice most of it. Do you have any hints as regards the a.a. composition of your protein? If so, you can measure the absorbance at 315 nm and calculate the protein concentration based on the % tryptophan content (molar extinction coefficients for this amino acid are readily found). Also read the absorbance at 280 nm: this will help you to find a tyrosine/tryptophan content ratio. Once you have found a sound value, you can calibrate your abs readings at 280 nm for future determination on the same protein.

My modest, helpful (albeit laborious) suggestion is: start from much more biological material and carry out your purification again and again to get much more protein.

All the best and good luck!

Giovanni Cercignani

One option is to run SDS-PAGE and use a sensitive protein stain such as colloidal Coomassie Blue or a fluorescent stain, alongside known quantities of a protein standard. If you have 10 ng/uL protein, you will only need to run a few microliters. This will not be exact, but at such a low concentration you don't have a lot of choices.

I dell with the similar task by the approach that Giovanni suggested - measure Abs280. advantage of this method is as such that you will not waste your sample (!) you can use standard spectrophotometer with 1 ml cuvette, or platereader with 200 ul. I personnaly used to detect as low as to 0.5 ugm/ml +/- 10%. Using this approach be sure that you use propper (quarz) cuvette or UV transparent plate. Good Luck

I've been able to use a Pierce BCA protein assay at below 25 ug/mL (typical lowest standard) while maintaining a linear standard curve (R2=0.9933). This requires 25 uL of sample per well. Note that it gives you non-recoverable protein (destructive technique) and only gives you total protein--it's nonspecific. I think I've seen modifications of the assay for nanodrop and ultra-low volume usage.

You can concentrate your material (5x) using millipore concentrator and then use BCA test kit to do the assay which can detect 10 ug/ml protein concentration. In order to avoid wasting too much protein scale down your BCA assay so that you have just 5 ul total volume. You can use 2 ul for measuring absorbance using NanoDrop apparatus.

I agree with Andrew Tai that you may be able to use SDS-PAGE with a sensitive stain. A fluorescent stain (See Life Technologies Molecular Probes catalog, for example) is probably more sensitive than colloidal Coomassie Blue if you have suitable equipment for illumination and quantitation. Use a standard of similar molecular weight for comparison. You don't necessarily even have to use the gel. You might be able to use the stain in solution if you have a fluorescence plate reader. If you use low-volume 384-well black plates, a fluorescence assay can be done in as little as 4 or 5 microliters. However, you have to be careful to avoid adsorption of the protein to the plate surface. This can be done by including a low concentration of non-ionic detergent, such as 0.005% Triton X-100. Check out this web page: http://www.invitrogen.com/site/us/en/home/References/Molecular-Probes-The-Handbook/Protein-Detection-and-Proteomics-Technology/Quantitation-and-Selective-Purification-of-Proteins-in-Solution.html

Thank you very much for your responses:

I have found 2 Pierce BCA protocols:

The regular one: http://www.piercenet.com/instructions/2161296.pdf

And the Micro BCA: http://www.piercenet.com/instructions/2160412.pdf

Micro BCA suits me in terms of sensitivity but requires at least 150 ul protein...

As for the regular one there are a few options - If I use the test tube procedure I have to sacrifice 100 ul protein and if I go the microplate way I can go as low as 25 ul per sample, which is not bad but I don't understand which one of the above protocols you (Khawar) suggest to scale down to 5 ul, and do you mean to use NnoDrop to read at 542 nm?

Boris - I understand you did not use NanoDrop as its lower sensitivity limit is 0.1 mg/ml...

We found in our laboratory that the best way to measure protein concentration (regarding sensitivity and accuracy) is to use spectrophotometric method. We use a cuvette that needs 70 uL, but since the method is not destructive to the sample you can recover it. (There are even some cuvettes that need less.) The important thing is that when you measure small volume you still keep a 1 cm pathlength. If you measure in a microtiter plate the pathlength gets very small and decreases the sensitivity since absorbance is proportional to concentration and pathlength (Lambert-Beer law).

We measure absorbance at 228.5 nm and 234.5 nm. Protein concentration is determined according to formula: γ (protein) = (A228.5nm - A234.5nm)×0.3175 [mg/mL]

It is much more sensitive than measuring at 280 nm which is expected if you consider the absorption spectra for proteins.

Reference: Ehresmann, B., Imbault, P., Weil, J.H., 1973. Spectrophotometric determination of protein concentration in cell extracts containing tRNA’s and rRNA’s. Anal. Biochem. 54, 454–463.

The method suggested by Marija is a very good choice; I was uncertain about the wavelengths, that's why I didn't tell in my first comment (I was at home in the early morning). That's the spectral region where peptide bonds start to absorb, the coefficient are hardly affected by secondary structures and disulphides scarcely interfere. Beware of using non-absorbing buffers.

my friend , I think u shud go with Lowry''s Method ,though its a destructive method but in my suggestion it vl gv u better results about ur accuracy for protein conc. U shud measure the conc @ 750nm wave length . u can prepare ur samples in diffrent dilutions such as 10 ,20 ,30, 40 ,50 microlitre or for more acuuuracy of result u make only 4 dilutions of 25 microlitre (25SSample+75Distlled Water) n multiply ur final result with ur dilution factor ie 4 (as 25*4=100 ).if u feel any difficulty b4 ur experiment plz mail me on [email protected]

goodluck

Warm Regards

@ Stephanie's post

NanoDrop indeed uses less than a microliter but you have to have in mind that the pathlength is around 0.5 mm and this decreases sensitivity!!!!!! (Lambert-Beer law) Detection limit for NanoDrop is around 120 ug/mL of BSA.

@ Marija - You are completely accurate regarding the effects of pathlength- sorry for being unclear, I completely left this out of the discussion (tried to edit, then accidentally deleted..ugh).

The suggestion would be to use nanodrop in conjunction with a stain (Bradford) - the bottom of the linear range, even with the smaller pathlength drops to around 15ug/ml which is close to the 10ug/ml original concentration (maybe slightly concentrate to ensure you're in the dynamic range). While you would effectively loose that sample, the low sample volumes would ensure that that amount is very small.

Hey guys(and girls),

Thanks a lot for all the great advises. I have tried following @Marija suggestion with the spectrophotometer but the deviations were way too high and thus the data didn't make any sense both for the standard curve and the samples - This could probably be the result of an old device.

Then I went forward to performing Bradford test with NanoDrop. Though the volume used here was indeed very low again the values I got did not seem to be very consistent leaving me no choice but to address ELISA.

Since I wanted to quantify affinity purified IgY I coated a 96 ELISA plate with 50 ul of the eluate diluted by 5, then blocked with 5% milk and reacted with anti IgY HRP. This finally worked and I am attaching the excel chart of the standard for you to see the linear range. My samples were 0.5-5 ug/ml so this is probably what made the other techniques fail.... eventually I used only 10ul*2 for each sample so it is not so bad.

Thanks again for all the good advice.

try OPA (sigma or Fisher Scientific) if you have any question don't hesitate to ask [email address deleted] .

You can also try Flurescin method with ELISA strip for which you just need to spend tiny amount of your protein saple.

Do you have access to a NanoDrop (Thermo Scientific). This instrument can read 280nm (assuming your protein is pure) with only 1µl sample

hi everyone, according to my experience (i've used: µBCA, Nanoorange,Coumassie, Fluoroprofile, ELISA simple and sandwitch, HPLC), the best remains ELISA sandwitch, u can detect picogrammes so in sensitivity it's the best, but the accuracy has a limit it depend of your protein and the AC u use, sometimes u can obtain an overesimation of your quantity. Try also fluoroprofile (need only 50µL and u can dilute your samples), keep in mind that's important to combine either a chemical and biological method to be sure. If u have any question i'll be glad to help u

Good luck.

I think this link is very useful, 

http://www.instanano.com/characterization/theoretical/concentration-calculation-from-uv-vis-absorbance/