In addition to all the above good suggestions, you could identify whether the two bands that co-purify with your protein of interest are contaminants from the e.coli strain that you are using by transforming the cell line with an empty vector (same vector but without your gene of interest) and treating it as a control. If you see same bands appear then that confirms the source. From here, you could use a different tag other than his-tag or double tag (his tag and another tag). On the other hand, assuming that the proteins that you see are truncated versions of your protein and that they are not due to degradation, you could solve the issue by simply moving the tag from N- to C- terminus. If it is due to degradation, you could add protease inhibitors during purification. That will hopefully solve the problem.

Some times proteins can be glycosylated or deglycosylated that could affect the effective mass of the protein and that being reflected in the SDS-PAGE run. The term you used " two light bands which are very close to each other and are pretty close to Hsp90 in the gel"---Are those bands appear in higher mol wt region than what is expected? Then I will suggest trying a different antibody and if possible try deglycosylation and try running on SDS-PAGE and do the western analysis.

Try adding 500mM Nacl in the Ni-NTA purification procedure. I believe they are interacting partners of your protein of interest. Increasing NaCl concentration in the Ni-NTA purification will not affect your quality of purification, but will remove non-specific proteins that are binding to your protein. If this does not work, do not hesitate to increase the salt concentration to 1M. I hope this will help.

Yes it's common to get non specific bands (proteins) along with the protein of interest after Ni-NTA purification. Which probably is due to the presence of repeats of His residues or the non-specific proteins have compatibility (direct physical interactors) with protein of interest.

Agree with Amit, increasing salt conc. would wash out the interactors. If problem still persists check out ur SDS page gel using the elution buffer in sample buffer in some the wells.

Additional to all this good advise, I have one very simple-to-check suggestion: When you are doing SDS-PAGE/Western Blots, load 1 lane with your SDS sample solution only. Sometimes "bands" can appear and make an evaluation impossible..

I hope you have checked and confirmed that one of those bands is your protein of interest with your protein specific mono clonal antibody?

There might be many reasons (you can have a look at the previous discussions in the RG). As said about increasing NaCl is a key thing, at the same time some hydrophobic proteins like to precipitate when the NaCl is higher, for such proteins I got success with 50mM to elute them (irrespective of washing).

Why do you need IEX step? is it due to the impurities or to separate these doublets? if you can not separate them with IEX, doing SEC is pointless.

You can make a trick, connect cation and anion columns in tandem, and pass the protein solution, and disconnect for separate washing and eltuion, this may give a different result! (worked in my case).

Why not to try buffers other than phosphate buffers(I hope you made the P-buffer by mixing monosodium and disodium phosphate solutions!)? HEPES/Tris-Cl are good alternatives.

Good luck!

Could it be caused by proteins in differently oxidized forms? These often appear on gels or westerns as "doublets". If so, try increasing the amount of beta mercaptoethanol in your loading buffer.

Bhaskar: You can attach the gel for people to help you better!

Would you like to cut the suspect band, get the sequence and blast to determine the nature of the protein? Western blot analysis could also determine whether or not it is related to the target proteins.

This happened to my protein actually. We used buffers at pH 7.8 and the protein came out non pure. I suggest you raise the pH to 8, because maybe what's happening is that the protein is not sticking well and could be eluting with these contaminants.

I guess an easy explanation would be that your protein degrades a bit during the purification. Determine the total mass of the proteins in your sample. The biggest mass should match your expressed protein, and for the smaller masses you can use the 'findpept' program on the internet to check if it matches with a smaller fragment of your protein. A bit of degradation would perfectly explain why the smaller bands co-purify until the end, because they will essentially have the same properties as the full-length protein.

Do you boil your sample before loading on gel? if your protein contains Asp-Pro bonds they will get hydrolysed during boiling. This will generate 2 or more polypeptides (depending on the number of Asp-pro bonds) during boiling, giving you the feeling that it is a contaminant, but actually is product of your protein hydrolysis!!

I checked on human Hsp90 and there is a Asp-Pro bond at residues 66-67.

How to solve this? Just load you protein with loading-dye BUT WITHOUT boiling!

Try. I hope it will work!

Did you try native PAGE without sds? Band of this gel may be help to you to make final decision.

As other people have suggested above, it may be glycosylation. Also if the proteins are lower molecular weight it may be degradation

and if the are higher molecular weight it may be aggregation. Sometime it happens that some proteins form homodimers or heterodimers. One other things you can try and do is increase the imidazole concentration during your binding and washing steps. This can help with getting rid of nonspecific binding (but not too much to lose your protein, up to 45 mM perhaps)

ALL THE BEST

Maybe, it is a truncated protein. Check the open reading frame (ORF) in you expression vector. Because, if you vector has two ORF, maybe you have a truncated protein and the full protein.

Just putting a couple quick cents in to echo some of Riaan's thoughts. When Mass Spec is done to protein fragments, very often there will be oxidized methionine. Multiply such modifications and you will see proteins give blurry bands or doublets which are all positive for your protein of interest on Western.

To me it looks like the main problem is to purify your protein rather than to understand why reason those other proteins appears close to your main protein. In this case, I recommend you to use an IEF-based purification method. I am pretty sure that there is enough difference in isoelectric point between them that allow their purification.

I hope this advice could help you.

They may be your protein isoforms. You can confirm them by sequencing if you want.

Following comments of Daniela De Biase and Yonca Duman: Use Laemmli loading buffer without DTT or bME as something between denature and native gel separation. Of course, no boiling.

There may be two subunit of protein or there is another protein (impurity) also. Try to re-purify it.

In addition to all the above good suggestions, you could identify whether the two bands that co-purify with your protein of interest are contaminants from the e.coli strain that you are using by transforming the cell line with an empty vector (same vector but without your gene of interest) and treating it as a control. If you see same bands appear then that confirms the source. From here, you could use a different tag other than his-tag or double tag (his tag and another tag). On the other hand, assuming that the proteins that you see are truncated versions of your protein and that they are not due to degradation, you could solve the issue by simply moving the tag from N- to C- terminus. If it is due to degradation, you could add protease inhibitors during purification. That will hopefully solve the problem.

Use the Novex loading buffer with SLS instead of SDS. Then you only need to heat your sample to 70. But I agree, why worry about what they are just go for the purification. You can try growing an E. coli strain that has reduced his binding

Have you tried to separate them on hydrophobic interaction column? For analysis sake, (if you managed to get only this two species, scrap off other contaminants) you can applied them to RP-HPLC and see how different their retention time. This approach worked to separate yeast's alpha- and gluco-amylases, which always co-purify to one to another preparation.

I agree with Douglas's suggestion. Maybe the protein degraded into 2 parts, and maybe the contamination of a Ecoli system.

They may be your protein isoforms. Increasing NaCl concentration to omit nonspecific band.

Probably it may be isoforms of your protien, check another pure protien to run on sds page, if your methord have some impure components it will clear. If nothing happends sequence all bands of your experiment.

While purifying, your protein might be getting degraded.....please look into trouble shooting

Hope you have run marker or a commercially available pure protein along with your sample?? there might be problems with the gel you prepare...!!!

Prepare the gel previous evening and allow it to mature for at least 12 hrs.......before you run electrophoresis

Cut the bands and try to do MALDi to identify if they are isoforms of the same protein.

Two bands may also be present because of protein degradation into two parts or subunit dissociation (if protein is multi-unit).

Your description is a bit unclear. The most important question is: can you see these 2 bands on anti-HIS western blot?If you cannot see them, they are contaminants. If you can-they are truncated forms (this also depends on whether your tag is on C- or N-terminus).

Contaminants may be removed by adding things like Triton (0.5-1%) or glycerol (~5%). But people suggested that already.

there must be the contamination in the protein sample,you must review your procedure,increase the NaCl concentration,that can remove the irrelevant proteins bands.

My first suggestion would be a western, probed with anti HSP90. If you still see the unwanted bands, it's protein degradation or maybe posttranslational modification. If they disappear, there might be stuff being attached to your HSP (HSPs work as chaperones, so this might be obvious) or proteins from E. coli that are co-purifying. Washing your protein bound to the Ni-column with a NaCl gradient (as already suggested) might do the trick.

Next thing to try is to add orthogonal purification steps: Fractionation with ammonium sulfate, IEX, HIC, changing the order of the steps. If you have an antibody against your HSP in sufficient quantities, an antibody affinity column is another option.

In my opinion check the isoelectric point of your protein doing two dimentional electrophoresis or IEF.

Regardless of the source of these contaminants, it seems you have useed the fractionated peak from anion exchange that later is concentrated by GFC as a sample to be loaded on SDS-PAGE,so you need to use an efficient method to get sharper peak. use pH gradient in anione exchange, or use longer column with smaller particle size.

or simply collect (fractionate) the target peak at above half of peak hight.

if you find out that the pI value of target protein and those impurities are two close, you need affinity chromatography to purify it.

to see what are those impurities, you can use MALDI-TOF mass.

It looks like your protein of interest is phosphorylated thats why giving you two closed bands, try to perform dephosphorylation assay on your protein to see whether this two band collapse to one band which will confirm its phosphorylation.

I am repeating the protocol once more with some minor modifications, would attach the gel picture. Antibody(anti-his)is something i would do as well. thank you everyone.

Another thought: There is the possibility of an analytical artifact: Boiling your sample in Laemmli "the standard way" might lead to degradation at prone sites. You might try lower boiling temp, shorter boiling time or modyfying the pH of your Lammli-style buffer. This easily may be checked using a (gradient) thermocycler. Replacing Tris by phosphate (the pH will change less with temperature) is another option.

Dear Bashar,

I suggest to consider the reducing / oxidant condiction of your buffers.

Try to add 1 mM DTT in lysis buffer, it will help to reduce the Cys of your protein which often bind other proteins. In my experience I used 1 mM DTT even in dyalisis, at the end of purification in order to enprove purification of my protein. Moreover add 1ul of fresh betaMercaptoethanol before loading the gel, sometime faster proteins are isoforms with many Cys-Cys links. Just to know, PBS precipitates when boiled. If you like to change your SDS-PAGE try the NU_PAGE invitrogen using Mes (6.5) or Mops(7.5) running buffer systems. In these buffers some proteins change their mobility on PAGE.

have good time

This is my SDS GEL picture after Ni-NTA, Anion Exchange and Size Exclusion Chromatography..kindly comment! Thanks to all of you above

The purification is not so bad, the point is what should you do with this protein?

I guess a lot has been said/suggested already. I agree with paola that your purification isnt bad at all. Occasionally such an artefact is due to re-oxidation of sulfhydryls leading to at some of the internal disulfide re-forming. This leads to a marginally reduced hydrodynamic radius and slightly greater mobility. This does look likely as the bands are equidistant. If you look up the method for counting disulfides, youll come across a very similar picture. I am quite sure that a tandem MS analysis will prove that all three could be the same protein.

It seems too simple an explanation, but you'd be surprised how often this happens. I would also like you consult www.aesociety.org (official website of the American Electrophoresis Society) for your problem. Very often the DTT or mercaptoethanol isnt fully reduced itself and a tributyl phosphine treatment (caution, extremely explosive) makes it better.

Best of luck. Do write back when something works. We all would like to continue learning.

Are you putting sample buffer in all of your lanes or just the ones with the samples? Dare I suggest the dreaded "fingerprint bands" might have contaminated your sample buffer? If so, you do not have a purification problem, just a gel problem. However if you have sample buffer in all lanes, that is not the answer.

Best wishes,

You do not have fingerprint bands. As is evident from the photograph. No problem with the gel at all. In fact i'd give it 8/10

I think you have a protein which is made up of three closely related polypeptides or you have three similar proteins of slightly different mw and in different quanta. The minor bands could arise from alkylated disulphides between the polypeptide chains. I suggest you analyse each band with MS and other techniques and compare your peptides from each band with one another and with other known proteins using an appropriate database search.

I wish you best of luck.

I also suggest you do a native PAGE of your purified protein.

The purification is actually quite good! But if your intended use does not tolerate having those two bands and if you need only a small amount of it, you could try the following: Run a gel with very many lanes of your sample. Cut the bands of interest out of the gel with a clean disposable scalpel. Electro-elute and then concentrate the protein down.

The additional bends can be proteins that contain a lot of histidines and tryptophan. which are known to adhere to nickel column.

I agree with western blotting and immunodetection to confirm the origin of your light bands

phosphorylation? or any other posttranslation modification such as protein carbonyls. That makes it slightly more heavy.

Many good answers: lots of histidine containing proteins in cell extracts but subsequent purification would have been expected to remove those. Moving the his -tag to the other end (COOH or NH3)is always a good idea when you have possible degradation.These are os specific size and possibly you have some normal cleavage going on. As Douglas suggested,putting a different affinity tag on each terminus helps- this is tandem affinity purification, or TAP, and is quite effective.Also as Musa suggests, before recloning try cutting out all 3 bands, digest and do MS on the fragments. A Western is sure to show they are all the same protein in 3 forms (phosphorylated or other modification as suggested by Helen,

The reason of dublet bands (peaks) is rate-limited self association. It is known and publish by Yang Shunong, Mariola Wolska-Klis and John R. Cann in Analytical Biochemistry 196, 192-198 (1991). The same macromolecule may appear as close dublet on the SDS-gels as well during the HPLC size exclusion chromatography. It is just pure physics.

Do any of these bands appear when you prepare a blank sample (reagents, but no sample) on your gels? If so, you may have a protein contaminant in one of the reagents. Also, if you are using reuseable glassware or plasticware for any of your solutions, you might have protein bound to the surface that is being eluted by reagents. One of the biggest sources of "phantom" bands is serum. So, if anyone in your lab is using serum for cell culture and the bottles are being reused for gel reagents, it could be serum proteins from these bottles.

It may be due to non specific binding of protein.

A western blot would probably give an idea of the source......but I believe this could be a degradation product of a protease action.

There are three reasons to get more than one band in western blot:

1. The protein may be contaminated with other proteins that have a very close molecular weight. so, actually more than one antibody is produced when you inject the protein into an animal for antibody production.

2. Your protein may be degraded during handling due to repeated freezing and thawing or temperature shift. I had purified a 49 kDa protein from mycybacteria which was easily getting degraded to 47 kDa by rough handing (purifying protein from crude extract without any ice).

3. The antibody may recognize another protein which may have at least structural resemblance with your protein of interest.

Another thing you can try is to improve your His-tag purification .

If you know your protein elutes at 100 mM imidazole, you can set up a wash at 80 or 85  mM Imidazole then run a gradient. This should give you a cleaner sample.

In your place I would also try with a non-phosphate based buffer.

I agree with above good suggestions given

the reason may be the degradation of protein samples during the  purification steps .

Hi, I have faced the same contamination problem as some E.coli protein co-purify with the your tagged His-tagged protein. Try Co2+ bead (from pierce) it is more specific than Ni2+ bead. I have tried it personally it works much better than Ni2+. One thing I would like to highlight that yield may get compromise as compare to Ni2+ but that you can manage by purifying with more induced culture. 

Hope it work for you.

There are lot of comments, I couldn't go through them. But basic questions I get are, do these bands show up in WB with polyclonal Ab too? If so, they are products of degradation, but they shudnt show up everytime in different extractions. Secondly, if they are interacting as nonspecific contaminants, I wud suggest you to use higher conc. of imidazole in ur binding buffer during Ni-nta step.

In addition to suggestions above, there are two more important points before the purification: 1) Expression of mammalian protein in bacteria needs codon optimization because the E.coli Bl21DE3 cells do not provide all or enough tRNA (anticodons) for codons on mammalian mRNA sequence. Finally, smaller recombinant protein fragments may appear on gel. For this situation, Rosetta E. coli strains may be used. 2) The optimum expression temperature should be checked. If the two bands close each other (doublet) was because of degradation, the expression of recombinant protein at a lower temparature can prevent this.