Desalting eluent fractions

[MestizoJoe]

Hi,

I have separatated, micrograms at a time, some components in a mobile phase that consists of octane sulfonic acid sodium salt, and potassium phosphate monobasic/dibasic (pH 6).

Would anyone be willing to suggest a desalting technique? Or how about a book or section in a book that might describe something like this.

Thank you.

[danko]

If your target is a larger molecule – like protein or something – you might as well run your fraction of interest on a SEC column. The target will then end up dissolved in whatever mobile phase you choose, whilst the current additives will be eluted with the total inclusion volume (i.e. somewhat later than the target). The technique is sometimes called “buffer exchangeâ€

[MestizoJoe]

Thank you for the suggestion.

It is a large molecule. An analyst I am working with is trying to desalt using a membrane that is supposed to allow molecules smaller than 3K daltons through. We'll see how it goes.

We will have to scale up soon so I believe i will take your suggestion.

[Alera]

Working with proteins, a cenrifugal filter MW cutoff membrane is probably a better way to go than GPC - it will also allow you to concentrate the sample at the same time (along wtih buffer exchange), if needed. It also alleviates the problem of having to inject a fairly large volume from your collected fractions at a time.

I used to use Amicon products a while back and they worked very nicely, especially the ones that were designed for the swing angle rotors, as they never go dry (if you happen to miscalculate the centrifugation time, which happens pretty often with protein solutions depending on viscosity of the sample).

[Gerhard Kratz]

I would follow Danko`s idea and go for a desalting step with SEC. If you want to scale up, what does it means? How much sample volumn you want to desalt in one shot? If it goes up to production size SEC is the better choice than membrans.

[HW Mueller]

Alera, the only amicon products I ever saw that don´t run dry are the ones recommended for rotors with a permanent angle. Due to the angle some liquid retentate stays on the corner of these filter units where there is no filter membrane.
With a detergent in there one might need some very good washes. Seems to me I have seen some special methods to do this, maybe in the Pierce catalog?

[Alera]

HW, I am sorry, that is exactly what I meant (brain freeze) - "FIXED angle" rotors. I was just looking on the their website and the only ones I found they carry now are the ones for the swing bucket centrifuges.

And you are right, I used to do 2-3 washes (depends on the residual volume and the volume you add, so you can calculate the dilution achieved every wash step). It is fairly easy to do - you just add your fraction to the filter, start the centrifuge, walk away, come back, add more buffer start it again etc. Recovery was good, too. And the material is concentrated after the last step, which is a bonus for a small amount of protein.

I am just not sure if you can inject a 1-3 mL fraction on a GPC column - it's not going to work - and it's going to be even more diluted after that.

[mardexis]

Is the octanesulfonic acid used as an ionpairing reagent for the species you are separating? If it clings to your molecule I'm not sure it will just fall off and through a membrane. Can you separate without it?

[Alera]

It is just a hydrophobic interaction, so if you are diluting your sample it is not going to "stick" to the protein, there will be an equilirium with solution.
I have done it many times.

[MestizoJoe]

To answer Gerhard's question, we have been using an analytical column to isolate preparative amounts of 6 components we are interested in.

Each component makes up ~5-10% of the total mass injected each time; ~700 ug if we are lucky. Plus we imagine we have some loss upon isolation. Our fractions after several injections and collections are about 5 mL which we centri-vac. We have reconstituted the sample and desalted with the membrane.

Soon we will scale up our separation by using a larger column, which will of course give us larger fractions and more salt. We will probably centri-vap the fractions and reconstitute them and try desalting by using SEC.


Question for Alera: It seems you would not recommend SEC for desalting a small sample. Is this correct? If so, if I were to scale up my separation and fraction collecting by using a prep or semi-prep column, and I evaporate the solvent in the fraction and reconstitute, making the sample more concentrated, would it then be a better idea to use SEC?

I'm hoping we will be able to load 5 mg per injection which might give us 50 ug per fraction (in about 1 mL). Our final fraction after 10 injections I hope would be something like 0.5 mg which can be reconstituted in 1 mL of solution (probably water or SEC mobile phase).

Thank you for your help.

[ScottHorn]

We do buffer exchange (desalting) on a daily basis in my lab. The products are large proteins (+100kd, up to 5000kd). Our method of choice is SEC using Sephadex. There is a whole range of Sephadex media, so I'm sure you can find one that works for the size of the molecule you're working with. I've found it to be quicker than the repeated centrifugation method, Sephadex media is relatively cheap, and can be reused a number of times. You can get satisfactory results just by pouring a simple column in a pipette with a plug of cotton in the bottom. I do sometimes use membranes to concentrate the sample before the SEC step to reduce the media volume necessary, or afterwards, as the SEC step usually doubles the volume of the sample.

[Alera]

MestizoJoe, look at it from the efficiency and ease of prep point of view. After you collect your fractions (about 5 mLs), you have to centri-vap them, which takes time and increases salt concentration significantly (potential issues with proteins which might denature at some point).

Using a membrane unit, you load all 5 mLs into the unit and spin it, concentrating and removing salt at the same time (salt concentration never increases). Once the volume is reduced sufficiently, you add your final buffer to dilute the sample and spin again, so if you spin twice, say 5mL to 0.5mLs, you will have only 1% of the salt left in the sample.

The reason I preferred this method over SEC was that it provides for less sample transfers (and therefore, losses), as sample concentration and buffer exchange occurs at the same time.

I used this procedure for fairly small to moderate (similar to your scaled up) sample amounts, but it sounds in your "scaled up" prep it should be fesible as well.

[skuhl9906]

Anybody have experience with centrifuge filter units NOT retaining the protein/analyte of interest? I'm working with a ~6kDa molecule and doing buffer exchange with the filter units (3k MWCO) the molecule is not being retained...

[ScottHorn]

I've never been in that situation, but there are a couple of things to keep in mind when using these membranes. First, the MWCO is not absolute. It's not like every pore in the membrane only lets through things smaller than the MWCO and everything smaller than the MWCO. I was told by a Millipore rep that a good rule of thumb is to have a two fold difference between the MWCO and the protein of interest (IE. use a 50kd or smaller membrane for a 150kd protein). Also, I *imagine* that the shape of your protein matters in the same way it matters for SEC. Perhaps your protein has a shape that allows it to squeeze through those pores. I'd go for a smaller MWCO.