Yes, 2 M Gu-HCl will affect your ion-exchange chromatography. You can get rid of Gu-HCl and all other low MW salts by dialysis against any low salt buffer with pH ~ 4.5, or just by Amicon UltraCel centrifugation cartridge with MW 3 kDa. You will remove all salts and concentrate your protein to desired range. For your ion-exchange step I would suggest something like CM-Sepharose (or any similar cation-exchange resin) with the low-salt loading/equilibration buffer around pH ~ 4.5. You can get your protein eluted in the buffer with the same pH containing 0.5 M NaCl. Just make sure that your loading buffer does not produce any salt precipitations with 0.5 M NaCl. Hope it will be helpful. 

Thank you for the advice. I would rather try to skip the buffer exchange and do this step via the ion exchange - which resin would I need to apply the sample in pH 8 and elute with pH 4.5? At the pI of my protein of 6.3, I suppose it must be a strong (pH-insensitive) anion exchanger, which is then unable to bind the positively charged protein during elution? Which resin type would you recommend?

Guanidine HCl, being a salt, will interfere with protein binding to an ion exchange column. This is also true for arginine, although the lower concentration of arginine may be acceptable. Try using urea (freshly dissolved) as a denaturant instead of GuHCl, since it is not ionic.

If you insist upon no buffer exchange, a guanidine-based refold, and the use of anion exchangers then there is a slight chance that a Q strong anion exchange fast flow resin may work at high pH near 10 -11 with significant feedstock dilution to make the binding as strong as possible in the presence of that much chloride conductivity from the refold. Unfortunately your protein is probably not stable to this pH. I believe Pall or Tosoh have an high salt tolerant anion exchanger.  Glycerol and lower temperatures may help lower ionic activity but there may be loss of binding capacity.  There may also be left over redox couples from the refold and misfolds that can adversely affect the concentrated product at the head of the column.  Keep in mind that switching the pH to 4.5 from an alkaline pH may cause on-column precipitation of misfolded product molecules as the column transitions to acidity, especially at the protein concentration at the column head.  A multi cycle batch-wise mode with diluted feed stock may be an approach to avoid column issues.  This anion exchange strategy is unlikely to work with cation exchange approaches although there may be a high salt tolerant cation exchanger as well.  HIC followed by anion exchange may be another strategy where greater purity may be achieved due to greater aggregate removal and the need for high salt for HIC binding.  

Sounds promising! Which would be a cheap and appropriate column material as strong anion exchanger? For laboratory scale, these are almost exclusively sold as columns, where may I acquire the free resin? Like Sepharose Q FF?

I recall Pall Corp. claiming to have a salt tolerant anion exchanger, maybe good up to 15 mS/cm room temp for the feed stock.  There should be a Pall sales representative for Germany so try calling the Pall office in Europe.  Most of GE Healthcare resins are sold as bulk resin because of manufacturing needs in the biotechnology industry.  GE also sells the bulk resins and empty columns so that smaller scale research operations and Process Development optimizations can be performed.  There's certainly a GE rep. in Germany so you can call the main office in Uppsala, Sweden to get price quotes for the Q FF.  Again, you might have to dilute the refold several fold with low conductivity buffer at high pH (maybe CAPS) or maybe you can use a Dowex anion exchange resin (OH form) to leave arginine and guanidine in the base form that will drive up the pH but the protein may also bind.  Once the protein binds to the Q resin then you can re-equilibrate at the your pH 8 condition and then run the chromatography but you should be very much on the alert for precipitation in the transition pH's if your elution intention is drop to pH 4.5; be careful.  

Thank you so much, Grant!

I played around with the pI and pKs-data, and the situation looks like this:

Assuming I use Q FF, the matrix will be charged (+) whatever the pH.

situation at pH 8 mobile phase:

o mPrP (strongly -, pI=6,3) = binds, FINE

o TEV-Protease (+) (pI=8,8) = elutes, FINE

o Urea (uncharged) (substitute for GHCl) = no interaction, FINE

o Tris (substitute for phosphate, buffering substance, uncharged or +) = elutes, FINE

o Chloride (-) but low conc like Tris = interacts, but sub-critical conc, FINE

o Imidazole (uncharged or +), pI=7 = elutes, FINE

o Arg (mostly +, should be fine) = low interaction, elutes, FINE(?)

So, binding should be safe, or did I miss something?

then, situation at pH 4.5 mobile phase:

(citrate/phosphate-buffer - should I exchange buffers/pH via gradient here or in a step?)

o mPrP (strongly +, pI=6,3) = elutes, FINE

o Citrate (buffering substance, uncharged or +) = no interaction(?), elutes, FINE

o Phosphate (counter-ion of buffer, charged -) = PROBLEM??

o Chloride (-) = used in abundance (400 mM), should cancel out phosphate interaction??

o Potassium (+), counter-ion for chloride = no interaction, elutes, FINE

o Arg (mostly +, should be fine) = low interaction, elutes, FINE(?)

Is a gradient for pH exchange better and how fast should I run the column?

Precipitation while crossing the pI is imminent...

I am planning on using the free resin in batch mode and pouring it in a simple column for fast buffer exchange.

For a gradient however, I would need an FPLC system, which is not readily available. Ideas?

Thanks again for helping me working that out in detail!

• TEV protease may still slightly bind  at pH 8 due to smaller % of neg. molecules in equilibrium with neutral and pos. species but the conductivity will likely be sufficient to prevent binding.  
• Dimethyl urea or dimethyl sulfone (not sulfoxide) can be used if you're concerned about carbamylation
• Imidazole at pH 8 will have tau nitrogen(s) with free pair of electrons in p orbitals providing basic character but not donating to the quaternary amine of Q 
• Arginine could be helpful in controlling aggregation
• No guanidine helps so all systems "go" for binding.
• Citric has carboxyl functions so only neutral or negative with pKa's ~3, ~4-5, and ~6-7 as I recall; also carboxyl is a weak acid but phosphate a strong acid. Both will help in the step elution with the NaCl.
• Arginine from refold will be gone once the load stage is complered.  But may want to have some there along with some chaotrope to inhibit precipitation/aggregation.
• pH gradients are not straight forward to control, especially in this situation, and slower residence times are needed to develop the pH and allow equilibration of the bound solutes with the pH.  This will also allow plenty of time for  precipitation/aggregation to occur.  You may need chaotropes.
• Probably should elute in batch mode  to avoid an unfavorable or precipitating pH value forming at the interface/front between the two buffers in the column mode.  May need chaotrope.
• If running in column mode then use limit of linear flow rate from the pressure-flow curve provided by manufacturer to create rapid pH change, but may need chaotrope.
• After loading, might want to rinse the resin with a very low concentration (few millimolar) of the pH 8 load buffer so that your step elution with pH 4.5 buffer will be more easily/quickly changed to pH 4.5.  May need chaotrope.
• Don't need FPLC:  just a good peristaltic pump (adjustable occlusion) couple of beakers connected by glass U-tube or two-chamber plastic gradient former (BRL) if needing gradient, silicon Pt cured tubing or Sanipure (20-30 psi), mag stirrer for one beaker, feed tubing from buffer A to column, detector and strip chart recorder.  With batch loading there will not be much separation so fast flow and low bed height will be consistent in producing low resolution with fast buffer exchange with minimal pressure issues on Fast Flow resins.  Manually collect fractions according to recorder trace.    
• Don't be surprised if you see some cloudiness appear and persist or disappear as the material elutes in column mode.  

Thank you again so much, Grant! I have already ordered Q FF and will try it out as we worked this out and as you proposed, will definitely implement all your good advice!

How much resin should I use to make sure to bind all my protein from app. 200 ml refolding buffer (from flash dilution)? I expect app. 100 - 150 mg protein (from 1,5 liter E. coli culture). Is less or more volume of resin useful for achieving better binding/separation? In principle it is just about washing the protein into the new buffer, so separation efficiency is more or less unimportant as there are no real contaminants at this stage?

Assuming the application is not for cell-based or in vivo work but only physico-chemical analysis then the protein may be sufficiently pure after the affinity step, may still have endotoxin for example.  If x-ray crystallography were the purpose, for example,  then the crystallization step would be even further purification with contaminants remaining in the mother liquor.  What will likely be present then is some TEV protease, His tag, and product-related contaminants such as aggregates, misfolds, clips, oligomers, etc.  The anion exchanger alone may not be able to get rid of all these product -related contaminants but there will be some charge difference as well as hydrophobic differences between these forms.  

The suggestions to date were made on the basis that a step elution by an acidic pH change was desired for AEX chromatography and that the load solution would be high in conductivity.  So the principle of buffer exchange with no purification would indeed apply.  However, there may be some contaminants that flow through and some that might not elute.  I'm quite certain that you will not have 100% refold efficiency and 30% is doing quite well.  You may be significantly below 10% but with your protein it may be very different than conventional proteins.  

But now you chose not to use guanidineHCl for the refold chaotrope and if you would like to run the AEX conventionally then a salt gradient elution (maybe 50 mM NaCl to between 0.5 and 1M ) at pH 8 developed in 20 CV's (column volumes) at ~0.2 CV's/min residence time could be used.  The significance of this is that you not only have a chance to remove product-related contaminants but you also get an indication of proper protein folding.  In most cases, correctly folded proteins elute in ~1/3 of the this 20 CV total gradient volume as a relatively sharp peak.  This peak sharpness criteria is also true of other chromatography patterns including RP-HPLC and even band sharpness in non-reduced SDS-PAGE.  For conventional proteins, industrially we usually use a load ratio of ~ 20-40 g protein/L IEX resin but a lower ratio of  

Thanks again Grant for the great advice! I have one last issue, I replaced GHCl with tetramethylurea (TMU), to prevent carbamylation with urea.

My aggregation assay conditions for the prion are normally 4 M urea and pH 2. 8 M urea pH 2 prevents aggregation. TMU is twice as potent as a chaotrope as urea. So for my first tests, I used 4 M TMU (=8M urea), the protein precipitated during mixing with pure TMU, but could be redissolved so that NMR was possible. However, after 2 days the sample was completely precipitated, while in 8M urea the sample does not aggregate - is it maybe not that simple to replace urea with TMU as I thought? Are there properties I did not take into account and is Dimethylurea better suited?

The acidic pH may will cause changes in the oxido-reductive and conformation conditions that the PrP can experiences during disulfide shuffling (no reducing agent).  The true pH in pure liquid TMU and possibly in 4M TMU will not be the same as in aqueous conditions; Henderson Hasselbach equation is based upon a water solvent. The pH may not be 2.  It would seem that the TMU is not able to completely denature the prion as does 8M urea but perhaps DMU will be more effective.  Are you using CD/polarimetry to look for helix to sheet transitions?