The concentration of albumin in serum is very high, on the order of 500 micromolar, or roughly 30 mg/ml. Your 1L of 10% FBS therefore contained about 3 grams of BSA. You probably precipitated a large part of it with 50% ammonium sulfate, so your main problem is getting rid of it. In addition to the excellent ammonium sulfate fractionation idea, ion exchange chromatography may be useful, since it has high capacity and can use inexpensive resins like DEAE-cellulose. If you use ammonium sulfate first, you will have to get rid of the salt before the chromatography.

First image

Second image

the distortion could be due to the presence of high salt, (you can check by adding BaCl2 to your sample to see if Ammonium sulfate has been completely removed., if not continue the dialysis ..)

If I recall it correctly, even PBS contains lots of salt, so that could be the problem. The desalting chromatography would be better for in means of time and efficiency of removing the sulphate.

The FBS, no matter what it is (as long as it is low-molecular weight) should be removed during the precipitation and dialysis/desalting.

I donot see any salt problem with your PAGE gel, simply because you loaded too much. If you load 1/20 of what you did, you will see a beautiful gel.

The dialyses you describe is extensive and should remove all salt - it looks to me as if you have overloaded the gel. I aim for loading about 5-10 ug of protein in one lane for complex samples and 2-3ug of pure protein to get the best results.

1XPBS is 155 mM NaCl (Gibco). This should not be the problem. Agree with Ziguo Zhang: don´t overload the gel. Is the second image a Western? If yes, check your technique. In any case, for any PAGE, check your APS stock and quality of TEMED. Both pictures give bit of an impression that the gels were not perfectly/evenly polymerized.

His-Trap columns are straight forward to use (manual), but as long as all you can tell about your protein is that it is a approx 50 kDA (monomer), there is hardly anything one can suggest further. If you don´t wish to make the nature of the protein public, go and talk to the biochemists "next door" who do protein works on a regular basis and get advice there. This will save you lots of time and money.

Before you process whole sample (1 L) you should first do some trials (pilot experiments) on small aliquots. I prefer to use desalting with Sephadex G25 instead of dialysis for two reasons: much faster and complete removal (if you load 20-25% of sample volume compared to volume of matrix in the column-25mL on 100 mL column). You can also check conductivity of your sample after dialysis to see if it is the same as the PBS. And, as Ziguo Zhang said try to load diluted (10x, 20x) sample to gel. Also, I dont see a problem with imac chromatgraphy? does fbs means fetal bovine serum? if that is correct than all small molecules (which cause problems by stripping nickel ions) will be removaed by desalting and you will have some other proteins which shoud not bind to column. You can do this with any imac column.. best wishes.

Dear Jason,

I am also working in similar system by using HEK293T cells and my some of the proteins are also secreted. I expressed several proteins successfully with very easy and handy protocol. you dont need any FPLC system for it.

1. Use advanced DMEM medium and then you will need only 2-3% serum. But even in my protocol, FCS will not have problem, you can use 10% FCS.

2. Spnt with your protein, e.g. if you have 500 ml expressed media then add Tris-Hcl 50 mM, pH 7.5, Imidazole 20-30 mM (it inhibits unspecific proteins to bind with resin) and NaCl 500 mM (you can prepare stocks of tris, Imidazole and NaCl). now adjust pH to 7.5 and make up the volume to 1L.

3. Keep this media in 1L duron bottle. Now add 5 ml recharged/fresh Ni-NTA resin to the media and roll over the bottle at 4 degrees overnight, dont use magnetic stirrer.

4. Next day keep the bottle standing for 2-3 hours, so the resin will settle down at bottom (you can see it). Now decant approximately 800-900 ml supernatant from flask by inverting down it carefully.

5. remaining 100-150 ml will have your resin. Now pack this resin to empty columns (Propylene empty colums; Biorad or any other).

6. During packing is you want to accelerate it, (becuase sometimes only gravity packing is very slow) use mild suction at column outlet by peristaltic pump.

7. after packing remove air bubbles by stirring resin and wash with 10 volume of binding buffer (Tris-Hcl 50 mM, pH 7.5, Imidazole 20-30 mM, NaCl 500 mM).

8. elute your protein with Tris-Hcl 50 mM, pH 7.5, Imidazole 250 mM, NaCl 500 mM).

Hope it will be helpfull for you, if you need any further help, you are welcome to ask.

I think it´s a PBS problem. Try desalt your last sample before PAGE, you can try spin columns or amicon concentration in order to remove salt excess, you´ll see the difference.

I don't have too much to add to the above, but one thing you should be aware of is that FBS also contains some pretty substantial amounts of some high molecular weight proteins. That, plus the salt, may have an effect on how your gels run.

It appears the loading is high in the western blot causing the distortion.

Directly running this protein on Ni_NTA Column without concentration will give a lot of background and may not work, assuming the expression of target protein is low. However, the present sample (ammonium sulfate concentrated) might do well on Ni-NTA column. In similar conditions, I have concentrated culture medium by 10-20 fold and buffer exchanged using ultra filtration (Stirred cells from Millipore) or TFF followed by affinity binding on Ni-NTA with success. For some low expression proteins there can be still some BSA and other bands . Optimization of Ni-NTA column binding/elution steps can eliminate these bands. Adding an ion exchange or gel filtration column as a next step can get higher level purity.

Wow, thank you all for the kind replies. I'll do my best to answer your questions:

@Tomas - the PBS I am using is 1X from Lonza, and yes it's about 150mM. In your opinion would it be best to redissolve the protein in a different buffer, say Tris-HCl with a lower salt content? I need to retain biological activity of the protein, but I'd also need it to be stable over time for future (in-vivo) experiments.

@Ziguo, Katrine - Thanks for the advice. I have been thinking that the distortion in the lanes is caused by salt, not from too much protein. I will do a dilution of my sample to get a rough idea of how much protein I can load. BCA would also help, but I haven't purified the protein, so the BCA reading would be much higher since the sample contains contaminating proteins.

@Joachim - Yes the second image is a Western. I did this just to confirm that my protein was still present after precipitation. You can see that there is a lot of distortion at the top of this image, presumable because there is a lot of high molecular weight protein contaminants, but obviously the 6x-His antibody didn't detect these. My protein is actually a dimer, so running a reduced, denatured gel will give me a band at ~25 kDa. Sorry, I didn't label the ladder in this image.

@Nikola - I performed a pilot study of the precipitation on 50 mL of media. I confirmed presence of the protein by western blot, and got signal at 40, 50, and 60% but the signal from the 50% sample was much more intense. I will try to run diluted samples on the gel next time. Yes, FBS is fetal bovine serum (some call it FCS, fetal calf serum?). I need to find out at what level of saturation the FBS precipitates, which would definitely help me to remove it during fractionation.

@Berindra - Thank you for the protocol. Yes I have heard of this method before, it's essentially diluting the media with your binding buffer, before adding the media to the resin. If you've had success with this, then maybe I'll try it, but is does increase your volume again once you dilute. Once question I have is when you add the Tris-HCl, Imidizole, and NaCl to the media, are those concentrations for 500mL of media, or after you top off to 1L?

@Prof Vujcic - Yes, I agree that using a larger reservoir will help but I'm mainly concerned about the protein not binding Ni-NTA when my media has FBS. I've attached a native gel I ran previously to show how the protein doesn't bind (or is washed out during the wash fraction). In this case no precipitation or dialysis was performed on the media because I wanted to see if I could purify the protein directly from the supernatant. Lane 1 is crude supernant, Lane 2 is flow through fraction, Lane 3 is wash fraction, and Lanes 4-7 are the elution fractions. Wash buffer is 50mM NaH2PO4, 0.5M NaCl, 20mM imidizole, and elution buffer is the same as wash buffer but with 250mM. You can see that the protein is still present in my wash fraction. Maybe I need to decrease the amount of imidizole in the wash fraction?

@Matthew - The large band in the first image is in fact around 65 kDa, so yes it could be BSA. This completely makes sense with the gel image because BSA, being larger than my protein, should precipitate at a lower saturation than my protein right? My guess would be that the BSA can be taken out with a 30-40% saturation before I reach the 50% saturation required for my protein. This should help remove the "blob" I see on my gel. Does anyone know at what percent saturation BSA precipitates?

Jason

All good suggestions. I would use ammonia sulfate fractionation as suggested by Matthew to reduce the FBS. When you load your Ni column, you need salt and low concentration of imidazole to prevent non-specific binding of other proteins to the column, as suggested by Birendra, so that the capacity of the column won't drop significantly. Load your Ni column slowly with a peristaltic pump and don't overload the column. Wash the column with some predetermined amount of detergent (you have to test and determine the amount of detergent required) to remove non-specific bound proteins. Wash out the remaining detergent with buffer and then elute your protein with high concentration of imidazole. Good luck!

The concentration of albumin in serum is very high, on the order of 500 micromolar, or roughly 30 mg/ml. Your 1L of 10% FBS therefore contained about 3 grams of BSA. You probably precipitated a large part of it with 50% ammonium sulfate, so your main problem is getting rid of it. In addition to the excellent ammonium sulfate fractionation idea, ion exchange chromatography may be useful, since it has high capacity and can use inexpensive resins like DEAE-cellulose. If you use ammonium sulfate first, you will have to get rid of the salt before the chromatography.

Hi Jason. I think by now you already understand the reasons for most of the problems you experienced, thanks to many excellent comments from friends. I give you some more tips to help you in the next step, because the real problem of your project has not been surfaced yet. First, you don't have to worry about salt or BSA because these do not interface with Ni-NTA (immobilized metal affinity chromatography, IMAC) chromatography. The serum however contains several IMAC-binding proteins, and you will suffer from the contamination of these proteins. Presumably these are proteins that are relatively rich in surface-exposed His or that have metal-binding capacity. None of them are strong binder and you will be able to get rid of them by extensively washing the column with a buffer containing 20mM imidazole+300mM NaCl, before the elution of your protein. However, you will also lose your His-tagged protein in this washing step, sometimes significantly. People often say something like "wash extensively with x10 column volume of buffer". Well, what is important is the TIME of the washing step, not the volume; i.e., you have to finish the washing step as quick as possible. Use wide-diameter column or suction (as suggested by Birendra). This way you can minimize the dissociation of your precious protein from the resin (which is of course time-dependent process). Also, binding to Ni-NTA is extremely fast (most His-tagged proteins bind within 5min) so don't incubate/mix the resin too long. Another common problem is the leakage of the metal (Ni) from the resin during the chromatography. Fresh and active Ni-NTA resin looks pale blue, but If you notice that it had changed to whiter color after sample loading and washing, that means that the resin is partially losing the metal. This can happen by many reasons, the major one being air oxidation due to the mixing procedure. The AS precipitation is a great idea, not just to reduce the sample volume but to increase the concentration of your target protein. Hisx6 tag has relatively low affinity for Ni-NTA, so in order to achieve efficient binding you need to increase the concentration of your protein (because binding is a concentration-dependent process). If you are using Hisx6, I recommend switching to Hisx8 or even Hisx10.

You will makeup the volume to 1 Litre. means your media will be diluted 2 times. the binidng buffer is Tris 50mM, Imidazole 30 mM, NaCl 500 mM, pH 7.5. for elution add 250 mM imidazole in same buffer. This dilution didnt have any problem, becuase you will add resin to it that will selectively bind to your protein. it doesnt effect purification. Ammonium sulphate precipitation is not required, if you dont have tag in your protein and you want to purify proteins by other chromatographic techniques, e.g. ion exchange. then you need ammonium suphate precipitation and dialysis. it is streight forward protocol that I told you and I used it at least with 10 different proteins, and also my colleagues use the same. it is as easy as to purify His-tag protein from bacterial supernatants. We never had problem.

you can maybe try using a centricon with a specific MW threshold. this gets rid of most stuff below the 50KDa i suppose. Proteins with a higher molecular weight should can be taken care of by simple size-exclusion. you can then precipitate your protein using AS from that mix. It just looks like you have a bit of dirt left: could be caused by proteins that nonspecifically bind to your protein of interest, or ubiquitous/contaminant proteins like keratin. such proteins could still be present in the media, and FBS proteins would not be the only ensemble of proteins which may precipitate out with the AS. your cells secrete a bunch of other proteins, too. But your lanes do not seem to be too dirty, so you're close.

Maybe you can solve your problem before the purification. Try to reduce the percentage of serum in colture medium. You can carry out the expression at 0.5-1% of serum. There are also specific serum free colture medium for the expression of secreted recombinant protein available.

Check conductivity in your samples and compare to similar concentration of BSA in PBS. Having ammonium sulfate dilutions for a calibration curve may help as well. This will show you how much salt you have in the sample and if it affects your purification. Vivaspin columns are great for buffer exchange, concentration and MW exclusion as Benedict Tan suggested.

These suggestions are all very valuable, thank you for your help. It seems that the best way to reduce the volume of supernatant and concentrate is with AS, but cutting at different percentages will also help to remove unwanted precipitants. @ Matthew - since my protein precipitates at 50%, and there seems to be quite a large band from serum contaminants, I can deduce that a lot of the serum contaminants must precipitate before 50%. I will try to cut at 35-40% and this should do a decent job of taking out a lot of these contaminants.

I also have an idea that many of you alluded to, which involves the Amicon centrifugal filters. For this I can use a 30kDa cutoff to remove low-molecualr weight contaminants, and collect the media not passing through the filter. I can then do the opposite, spin with a larger cutoff filter like 60-70kDa, and this time collect the volume that passes through the filter. In this manner I've removed both low and high molecular weight proteins, but my only concern is how much of my protein I could lose on the filter of these devices. The column that Dr. Vujcic suggests may also be a great method to remove serum, depending on cost.

For salt removal (if this is even necessary after AS), the consensus seems to be to use a Sephadex G-25 or similar desalting column. I'm still a little torn on the Ni-NTA column purification, but I will try several of the suggestions including dilution of the sample prior to loading, speeding up the washing time, as well as using the Histrap Excel columns I ordered for direct purification from supernatant. I'm sure it may also require some optimization of the imidazole concentrations during the wash steps.

I'll try to post a gel image after Ni-NTA...

For the AS precipitation, you can try the sequential precipitation, but you should try it at first with small aliquot. For my first protein I have worked with, the activity was in all fractions from 20 to 80% and even in the supernatant after 80% precipitation was still some activity, so I rather precipitated only until 20% and used the rest for chromatography. For a new enzyme I'm working with right now, there is a nice partitioning so that I can first precipitate between 40 and 70% and only by that I get rid of about third of proteins from crude extract.

As for the filters, that is somewhat complicated. It's not exactly that 10 kDa cut-off pass all molecules smaller than 10 kDa and all bigger stay and for 70 kDa cut-off leaves everything smaller than 70 kDa and everything bigger stays in supernatant. Especially not all smaller proteins will pass, they will be only diluted but still stay partly in supernatant. So if you wanted to use them, you could use e.g. 30 and 100 kDa filters. I was thinking about something like that as well, but with the large cut-off you will have big loses. Better would be preparative size exclusion chromatography with resolution around 50 kDa. And SEC would be also faster.

The Blue affinity column is used for albumin removal, I've read about that although I haven't used it for this purpose. But also your protein could bind to Blue gel if it's supposed to bind ATP, FAD, NAD(P) etc.

The precipitate after AS contains salt, but it is questionable, whether you need to remove all the salt (e.g. for ionex) or not (e.g. for Ni-NTA).

Berindra - Thank you for the steps you described in your purification method. I attempted the Ni-NTA purification with 25 mL of harvested supernatant and followed your procedure, but I didn't have much luck. I've attached an image of the Coomassie-stained gel (lane 1: flow through, lane 2: wash, lanes 3-7 elution fractions). It looks like there is still plenty of contaminating proteins in the elution fractions (my protein should be ~26kDa after reducing/denaturing), so perhaps I need to increase the imidazole concentration in the binding buffer? In this case I used 20mM in binding, and 250mM in the elution. Thoughts?

Dear Jason,

Use 50 mM imidazole in binding buffer. I also used it with some proteins. I think you will be able to get better pure protein. If you see 1 and 3 all the proteins seems to interact with resin. it means they are unspecifically binding. If you increase imidazole in binding it can decrease contamination definitely.

Another point.

Now you can use advanced DMEM to get rid of FCS amount both media has almost same prices.

Hello friends,

So, high salt wasn't a problem at all, it was my mistake of overloading the gel :) After diluting the samples 1:10 through 1:50 in increments of 10, I was able to get a much better gel image. Thanks for the help!

@ Berindra - Your suggestions worked! I repeated your method using 50mM imidazole in my binding/wash buffer and the majority of the contaminants were removed. I attached a gel image. Still need to do western to confirm that the band is actually my protein, but it's in the correct range so this is a good sign. Thanks so much for your help. Now to perform a large scale purification...

@Matthew - I'm also trying your method of AS precipitation on a 1L culture. I salted out at 40% and discarded the pellet, then continued to 50% where my protein precipitates. I redissolved the pellet in binding buffer with 50mM imidazole, and I'm going to try a small scale purification of this concentrated sample without desalting using Ni-NTA resin. This will give a good indication of binding without desalting.

I am happy for your results. Good luck with your experiments.

Regards,

Hey guys, I know this post is starting to get long, but I just wanted to post some more results based on your suggestions, which have been very helpful btw. Matthew, I used your advice to remove the 40% fraction with AS precipitation on a 1L culture volume. I continued to 50%, and resuspended the pellet in 20mL of binding buffer (with 50mM imidazole). I took a 1mL sample of the resuspended pellet and added ~0.5mL of Ni-NTA resin to the suspension and basically followed Birendra's steps (incubating O/N, washing with binding buffer, then eluting with elution buffer with 250mM imidazole).

I also had previously precipitated a separate 1L volume at 50%, but instead of dissolving the pellet in binding buffer, I dissolved it in PBS then dialyzed further against PBS. The same purification steps described above were used on 1mL of this dialyzed sample.

After purification I wanted to compare both samples on SDS-PAGE (image 1: lanes 1, 2: elution fractions without using AS precipitation, lanes 3-6 elution fractions after AS precipitation; image 2: flow through, wash, and elution fractions after AS precipitation and dialysis). Lane 1 in the second image is obviously overloaded, my mistake. It's interesting to note that when I don't use AS precipitation and purify directly from the culture supernatant 50mM of imidazole seems to be enough to remove most contaminants during binding/washing. However, when AS precipitation is used there are still quite a few higher molecular weight contaminants. Increasing the imidazole concentration may help, but may also remove my protein. Maybe it's just best in this case to not use AS precipitation?

Second image

Dear Scientist.

Hope everything is going well with you. I am doing an experiment to identify a recombinant protein from the transfected cells (HEK-293 cells) by western blot. I can find the specific band from the cell lysates but not from the supernatant. The protein is secreted in the supernatant as I can detect it by ELISA.

Could you please give any suggestion, how can I detect it by western blot. 

@Md. Golzar Hossain 

Do you use the same antibody for both? Are the antibodies protein-specific or against tag?

Hello Golzar,

If you have tag (like his) in your protein then purify it. Some proteins also make insoluble bodies inside HEK cells, if they are not secreted. If you are not able to purify any protein from 500 ml supernatant and only get band in whole cells then it is definitely residing inside cells. There are other ways to purify those proteins.

Thank you Tomáš Hluska and Birendra Singh

Yes I use same antibody against tag. I always use same protocol for supernatant and cell lysates. I purify protein from cell lysates and then get the band but when I also purify protein from supernatant using same protocol and reagents I don't get the band. However after purification from supernatant the ELISA titre is usually higher then the data of the unpurified supernatant. 

Please if give me some suggestions regarding this.....

Birendra Singh

does your protocol about "purification of his tagged protein from hek293 condition media"  works with all nickel resin brands?

I work with His select affinity gel from sigma . does this protocol work for me?

Hello Nafiseh,

I am sure that this purification will work with all Ni-resins. The difference is that some Ni-resins has higher amount of Ni immobilized in comparison to others. All resins should work basically in a similar manner. Dont forget to add 30 mM imidazole in your binding buffer to get rid of the impurities. Second, When you transfect the cells, use minimum amount of the FCS. e.g.You can use Advanced DMEM media that need only 2% FCS during transfection and further change. Sometimes, high amount of FCS interefere during your protein purification.

@Md. Golzar Hossain

I meant whether you use the same antibody for Western and for ELISA or whether you have different antibodies.

I'd try to increase protein load and/or decrease antibody dilution to see whether you'll get other bands. It could be, that the antibody in ELISA reacts with something else (like with several protein species) which are in total present in higher concentration but each is in lower concentration so you do not see them on blot.

@Nafiseh Sanei ata-abadi and Birendra Singh 

I would avoid the imidazol in binding buffer (especially such high concentration as 30 mM) because some proteins may bind only weakly and thus may pass through with imidazol in binding buffer and one can always use several concentrations to see where the purity will be the highest.

Dear Thomas,

If you dont add any imidazole in the supernatant of HEK cells. There will be several impurities. The cell culture media has 2-5% FCS that have thousands of proteins can easily contaminate the purification. I agree that some proteins might elute at low concentrations of imidazole, but they could be only 5% proteins expressed. It can happen due to improper exposure of His tag. At first instance I always go with 30 mM imidazole in binding buffer. If it dont work then ofcourse it can be optimized. I did several purifications and never get any problem.

Dear,

Facing same problem as Jason White and got much information on this discussion. I want to know after using the resin as suggested by  Birendra Singh can we reuse the resin for next time.

Please suggest

@Fahad, yes you can reuse the resin. I've been using Thermo His Pur resin and after eluting you can wash the resin, then recondition it with MES buffer, then wash again, then store resin in 20% ethanol. I believe the protocol is on the Thermo website.

@Birendra - Have you tried this procedure with Protein G resin (adding the resin to filtered supernatant and shaking) to purify mouse IgG? I will need to replicate the Ni-NTA procedure with similar large volumes from a hybridoma I am getting, so I was curious if this protocol would work too. 

Or does anyone else have a tried and true method for Protein G purification of mouse IgG from hybridoma?

Jason

@white, I have used the protocol from @sing and got the purified protein but problem is that I got two band in SDS PAGE even tried to use the higher concentration of imidazole for washing step. 

@fahad I feel your frustration. Unfortunately this process takes a lot of empirical testing to figure out the right buffer, additives, and concentrations to get a pure protein from the ni-nta column. If you haven't tried already, try diluting your supernatant 1:1 with binding buffer containing 30-50 mM imidazole before putting it on the column. I also had high concentration of NaCl, like 0.5 M in my binding buffer. I use my binding and wash buffers interchangeably, so keep in mind if you do this and you're washing the column after binding, that your imidazole concentration is higher during the wash steps since it's not diluted (and you might lose protein during the wash). If your bands are far apart on sds-page you might have to do a second polishing step with a size-exclusion column to remove your contaminant. 

@white dear, I have already tried this as you described about ni-nta but no fruit. I am now doing ammonium sulfate precipitation followed by gel filtration in S 200 column and let see what happens. 

@ Singh I am also facing a problem in protein purification from HEK 293T supernatant but I never add any buffer into the culture  supernatant, I directly add Ni-NTA resin (after washing the resin in 20 mM of Tris -HCL ) and leave it tumbling for overnight then elute with 20 mM tris with 200 mM imidazole but I never see the band in SDS-PAGE. I just read the comment which you have given to White I am thinking to use the same protocol for my protein also and protein size approximately 55 kDa. or any other modification I have to do in buffer please let me know.