method carry over problem (very hydrophobic protein)

[miro2009]

Hello,
I've been trying to develop an RPHPLC method for analysis of a certain protein (rec. SAK) and having problems of carryover:

in details, after sample injection, I run 1 or 2 blank runs and still observe a peak at the same SAK retention time, it decreases from the 1st to the 2nd blank run... I validated that it's not an injector problem.

I'm using a C4 column, so I tried to go for a C3 column and polymeric RPC (polystyrene-divinyl benzene resin which as I'd read is of lower hydrophobicity of silica based resin), still the problem exists (slightly lower % of carryover but still unacceptable).
And also with all columns, I observe the % gets worse by time, ie., after injection of more samples, especially if they are supernatants and not purified protein.

I don't know what else to try, my mobile phase is
A:0.1% TFA in water, B: 0.08% TFA in acetonitrile.
I also tried changing TFA and used 0.2% formic acid, but obtained worse profiles (regarding peak shape, lower absorbance)

Can anyone suggest other approaches to these kind of problems?

[grzesiek]

"I validated that it's not an injector problem. " - how?

"slightly lower % of carryover but still unacceptable" - post results

[danko]

Hi Miro,

So, why do you think the hydrophobicity of the column constitutes the problem?
The fact that you’re using a C4 indicates that the problem lies elsewhere.
How about the mobile phase for instance? And in the other end of the line, what about the injector washing procedure (i.e. solvent, time, cycles, volume).

Best Regards

[miro2009]

Thanks Danko for your reply.
What do you mean by "Elsewhere", you mean a problem with the system itself?
One of the other reasons I narrowed down the possible causes to being a "method" problem and not a "system" problem is that we simply don't face this carryover problem with any other analysis methods of other proteins on the same system, even after injection of very high concentrations (but using other columns, C18 for instance)

In answer to your question, the injector is washed, after each run, with 200ul of 50% methanol which I think would remove any trace sample remaining on the needle... as for the mobile phase, what do you suggest?

[danko]

Hi again,

I’m not fully convinced that the washing liquid is optimal. In fact I would describe it as not optimal. You see; very few proteins would be dissolved in 50 % methanol/pure water mixture. So, it might be the first step you should take – test the solubility of your sample in the injector washing liquid.
The problem (if it was a problem which I believe) can be solved by adding some acid to the washing mixture (f. x. acetic acid).

Otherwise it could be the mobile phase that needs some more attention. But it’ll require more thinking and experimenting. So, maybe you should start with the washing and then go further from there.

Best Regards

[Eric Moore]

Hi, miro2009 -

I'll concur with Danko. I have experienced carryover that has been corrected by using a needle wash solution in which my protein has better solubility. In our case, we switched AWAY from 10% methanol (which worked well for other molecules, but in which our protein was not soluble) to 0.1N HCl.

EM

PS: If switching to acid, be sure to wash out the needle wash lines regularly.

[Uwe Neue]

There is a simple test (developed by my colleagues at Waters) that allows you to tell, if the carryover is from the column or from the instrument, such as the injector. The way it works is as follows: you run your gradient twice through the column without using the injector. In some instruments, you can bypass the injector, in others you just need to program a double gradient run (with all the equilibration between the runs) into a single gradient table. If you see nothing in the second gradient, it is not the column. If you see something, the problem is with the column or the gradient.

[miro2009]

Thank you Neue for your input, I will try to perform the steps you described, because blank runs performed by a LC2010 CHT, ie. by inserting a -1 value as the vial number, the injector is still not by-passed. By-passing the injector has seperate settings from the instrument itself I will need to check.

But anyway, if it's an injector problem, shouldn't I be having this problem with all the other methods perfomed on the same system? this isn't the case

[Uwe Neue]

If it is an injector problem, it is not a mechanical issue, but a question of your analyte sticking to things like the seal material. This can be specific to analytes or analyte classes - such as hydrophobic proteins...

Anyway, the experiment is clean and will tell you, where to look next. The idea is to eventually find a solution.