Dialysis: how to avoid volume increase?

Emir Khatipov khatipovNO at NOuchicago.edu
Mon Mar 18 18:09:19 EST 2002

You can get PEG 15000 from Sigma. I remember in old times there was even PEG
20000. It might still be available from somewhere. Last time I used PEG
15000, I just dig the dialysis bag with the sample into the beaker or a tray
with PEG powder. Works very fast - 2-3 hours and you are done. You even
should be careful not to dry out your sample completely. I would use 3000 Da
or even lower cutoff membrane even with PEG 15000. As for impurities, after
you sucked off enough water from your sample, you can again dyalize it
against the buffer of your choice to get rid of those.

You can even remove liquid quantitatively if you wrap the dialysis tube
around one end of a graduated tube that you dig into PEG powder and follow
the decrease of the level of your sample over time.

If you protein is not too sticky, you may try concentrating it over ion
exchange resin. Bind at low salt, elute with high salt. I believe there are
monoQ and other concentrating cartridges available from Pharmacia, Waters,

As a last resort, there is always freeze-drying...


""Peter Cherepanov"" <peter.cherepanov at uz.kuleuven.ac.be> wrote in message
news:000901c1cebd$3af414e0$9293a8c0 at uz.kuleuven.ac.be...
> Does anyone know how to avoid increase in sample volumes during dialysis?
> Anything better then PEG (polyethylen glycol)????
> My problem is ethylen glycol (from "hydrophobic elution"), if I want to
> rid of it by dialysis, sample volume increases greatly, the dialysis bag
> swells to its maximum possible size, I even think it can explode if I
> more sample in it. Concentration of EG at start is around 35%.
> Now I am trying to dialyse first against PBS (it swells like creazy, I new
> it!), then PBS+10% PEG (still waiting to see it shrinking...)
> Another question - will there be a problem with PEG-8000 if I use membrane
> of 10k cut-off, any bad experiences?
> (PEG is a linear polymer, the reall cut-offs of the membranes should be
> smaller for such molecules then for globular proteins, at least it is what
> think ... ).
> Peter
> ---

More information about the Methods mailing list