Hi,

It is clear that the binding efficiency reduced because of the glycosylation.

Will you tell me the conc of Imidazole in wash buffer and elution buffer.

Buffer composition?

Find the enclosed purification handbook for further clarification.

Goodluck

i am not an expert of recombinant proteins but sometime during the purification even salt contration play a big role in protein binding and unbiding to particulars pritens and its type. Is it possible that you can try different naCl concentration in loading buffer? I hope it may give you some clues. Good luck.

Hi Christoph,

it is not neccessary the glycosylation, that prevents binding. We have encountered some problems with Ni-NTA binding with secreted proteins from insect cell cultures. We have observed, that some additive in the medium (in our case serum-free medium) interferes with the binding. We concentrate (ultrafiltration) the medium and change the buffer via PD-10 columns - then our (glycosylated protein) binds perfectly to the column. We also tried different resins (Ni-NTA Qiagen, Sigmas His-Select and Talon Cobalt resin, respectivery) - Talon cobalt worked best in our hands (try also to either apply your protein several times on the column or to apply it in the batch mode). We have a 8xHis Tag but also a 10xHis tag could be also worth to try. If you use serum in your media, there is a lot of - let us call it "dirt" - around, when you load your protein on the column - this is not the case, when you use intracellularly expressed proteins. There you have harvested the cells and discarded the medium.

Good luck. Best regards,

Magdalena

Sorry, i didn't mention it in my question: my protein is to large to pass the yeast cell wall and therefore i am not working with culture medium but cell pellets like i do in the case of the intracellular protein. I allready performed a ammoniumsulfate cut to get a purer starting material and than changed the buffer by extensive dialysis. That didn't help.

I tried batch purification from 1h to o/n, wihtout any success... I am using a 6x his tag, so your suggestions concering tag-length and resin type could work. Though at the moment i consider that to be some kind of last resort :) Thanks for your advice

Hi,

Glycosylation may result in aggregation of your protein or interfere with the His-tag binding to the resin.

According to my experience, I used 1-1.5M NaCl in Ni-NTA buffer to circumvent the interaction of my protein with DNA. If 8M urea had no effect , maybe you can try 1M NaCl, hoping it can decrease the electrostatic interaction.

If all these do not work, you may need to change your construct, say replace the C terminal his with an N terminal one or other tag.

Try to increase the pH of your binding buffer. That way you will increase the Ni-HisTag affinity. Change of pH may also cause a conformational change of your protein's structure, which may be in favor of stronger binding... or opposite. Anyway, it is a simple thing worth to try.

hey Christoph,

trying it once again after the site crashed in my browser....

So, based on the information you gave I have two questions:

A) Do both forms you express have the His-Tag on the C-terminus? You did not explicitly mention whether the intracellular form is expressed in the same way

B) Are you sure that the His-Tag is still present on the expressed protein? Proteolytic processing on both, the N- as well as the C-terminus can occur in the ER and golgi resulting in the loss of amino acid sequences that have been there in the intial translation.

How do you check for your protein in the flowthrough/binding fractions? using a HisTag antibody or with a specific method for your protein?

I don't think that glycosylation itself will interfere, however, based on what you mentioned, if your protein is really large and contains several glycosylation sites the protein will experience a significant mass increase that might lead to the protein being less well retained since the His-Ni interactions might not be strong enough to keep it bound.

Just some ideas....

Best

Daniel

Changes in pH should be help the stabilization of the coordination between Ni and His. In principle a deprotonation of the His ring by increasing pH has to be helpful. I recommend the article “The binding of Ni(II) ions to hexahistidine as a model system of the interaction between nickel and His-tagged proteins” (http://dx.doi.org/10.1016/j.jinorgbio.2005.11.003) for further details.

We have had similar issues with Pichia expressed proteins. There are components in the expression media and/or produced by Pichia that interfer with inding. Binding was improved by first dialysing the media. Using phosphate buffers, rather than Tris also improved binding. Pichia hyper glyosylate proteins, so you might consider Pichia GlycoSwitch SuperMan5 which yields proteins with shorter CHO chains ending in Man.

Hey Christoph,

as I understood your question, the protein is the same in both cases. The differences are 1. (to minor degree) the solvent and 2. the presence or absence of glycosylation. If, as I would assume, additives have no or only mino effects, the glycosylation is likely to influence the binding - either due to steric hindrance or due to secondary effects (aggregation...). In this case another resin is unlikely to improve the performance. If you are not interested in the glycosidic moiety itself, you might simply remove it enzymatically (PNGase A or PNGase F).

For the record, PNGase A will not work on the intact protein at all (it requires a proteolytic digest of the protein prior use), and PNGase F is usually working very poorly on native glycoproteins and requires some kind of denaturing process or detergents to efficiently deglycosylate a protein. Not sure whether that is an option in this case.

Tough problem, proteins can be resistant to denaturation by urea. Perhaps the His-tag is making an ionic interaction in the glycosylated form of the protein. Have you tried using 6M guanidinium to lyse the cells (or making the lysate 6M GuHCl) to denature the protein before binding to the resin? That should rule out/eliminate any steric issue from the glycosylation/altered protein structure in blocking the His-tag from binding. Just check for compatibility with your resin (can use it with Talon resin with caution). If the protein binds in 6M GuHCl, you can then do an on-column re-folding by exchanging the buffer (probably best done using a gradient at 5oC) before eluting the protein. I would also not use any imidazole in the lysis buffer until you had it binding good. And as mentioned by others, adding NaCl to 2 M could also help. Your anti-His-tag WB result tells you that the His-tag is accessible under denaturing conditions. Have you quantitatively compared that WB to one with the non-glycosylated protein to determine if all the glycosylated protein has the His-tag?

I agree with David Johnson. Dialyze of the media is very useful to remove contaminants. I dialyze medium against phosphate buffer 20 mM pH 8.0 and prior to pass through the column, I add NaCl and imidazole to obtain a 0.5 M and 20 mM buffer respectively. Also, to remove cellular fragments I filter the extracellular medium by 0.22 uM after centrifugation of the culture.

it is possible that the glycosylation is shielding the binding of the tag to the metal. Since, I believe, its high mannose its likely not charge repulsion rather steric hindrice or perhaps in the unfolded state the the tag is not accessible. perhaps the buffer exchange of the material to remove the denaturing agensts will not only remove chelators and other materials found in the media that might inhibity protein binding to the metal but also the change inconformation will result in accessibility of the tag for metal binding?

In addition what Knut and Daniel wrote - you can try some mannosidase or one of the Endo endoglycosidases (which should work, Pichia does not produce complex glycans). These work also on native glycoproteins, o, then you can check, if the binding to Ni-NTA is better ... Good luck.

Can you compare elution profiles of the glycosylated and non-glycosylated preps over a gel filtration column? If you're getting aggregation/multimerization you should get your glycosylated protein eluting much sooner. Did you try adding a little bit of detergent?

In principle Sergey is right, however the lectin approach will not specifically pull out your protein but a variety of different proteins all carrying the respective glycan epitopes.

I think a column suggested by Sergey followed by a gel filtration would probably help the purpose better.

Here is a demonstration of a protein purification technique, specially for highly glycosylated proteins.Hope this would help.

Can you just purify the protein by conA and other chromatographic methods like the old days when people did not purify recombinant proteins? I am just asking because I am old. We used ConA, ion exchange and another affinity step to purify TACE, a protein in low abundance, so I would think your expressed protein could be purified by similar techniques.

Never mind, I just saw Sergey's answer.

I did a search on glycosylated proteins and metal binding, it turns out that glycosylated proteins (like albumin in diabetic patients) can bind copper and iron (see: http://www.ncbi.nlm.nih.gov/pubmed/12162426). Maybe your glycosylated protein is binding Cu (which can also be used to bind His-tagged proteins) and either before, or during, loading on the column the His-tag on one protein binds to the bound Cu on another protein causing aggregation (as you were thinking). Extensive dialysis of the lysate before loading on the column, as suggested by Salvador, may solve the problem.

One way to test it would be to strip a column of the nickle, and then load the crude lysate onto it to see if the glycosylated protein with bound Cu binds.

Christopher and Gary, if that is the case, there is also something I think that is called chelex that you can treat your protein with to remove metals. It is used for enzymes when one wants to determine mole equivalents of cofactor metal to enzyme amount. The chelex is like EDTA except it is a resin. So I think the way it works is that the chelex is added to the dialysis container. This is worth a try, I think.

I think for your own satisfaction you should enzymatically de-glycosylate the glycosylated protein and then see if it still binds Ni-agarose. You can then confirm it is completely a glycosylation inhibition and move to other lectin-based purification methods.

You could always just purify both proteins using cation/anion exchange media. Not as straight forward as His-tag media (but then in this case it is not that anyway). The glycosylation should not affect the PI, so both proteins should bind/elute to cation/anion media under the same conditions. If they bind to cation media, then you can first batch bind the lysate to anion media to clean it up before loading onto the cation media. I have purified proteins to ~90% pure using this approach.