High pH mobile phase buffers

[Mattias]

I have struggled with a separation of 8 peptides for a period of time now, and have finally found the right conditions to get baseline separation of all peptides.

The pH turned out to be a very powerful parameter (many times more important than the packing material). The optimal separation was performed at pH 11.0 (works with both Zorbax Extend and Waters XTerra)I have never worked at this pH before, and I have some questions:

1. At this pH the peptides are unstable, but the peaks look nice with little tailing and good plate counts. How can I be sure that I don't get any on-column degradation?

2. I have so far worked with ammonium hydroxide solution pH adjusted with sulphuric acid to pH 11.0. I am a bit worried to use ammonia, since pH 11.0 is to far away from the pKa of ammonia (9.2). I am hesitant to use DEA and TEA buffers since they have the reputation to disturb the LC-MS analysis. Do you have any other buffer suggestions (compatible with LC-MS?).

3. Will my LC instrument survive in the long run when exposed to this pH? So far it looks OK, but I have heard that some parts in the instrument should be exchanged with more resistant materials.

Any other experiences of high pH separations??

Thanks!
/Mattias

[SIELC_Tech]

Mattias,

You don't need to go to high pH to retain and analyze your compounds. Check the following link. Our newsletter:"Is There a Mystery in HPLC Separations?" provides comparison to between mixed mode chromatography and analysis at high pH (Gemini, Xterra, etc.). You can see that our approach allows to retain compounds at much lower pH (2-7). Mixed mode chromatography allows you to retain a wider range of compounds (amines, amino acids, acids, peptides, etc.) Separation is easily controlled by ACN and amount/pH of the buffer;

http://allsep.com/brochures/SielcNewsletter_0605.pdf

[SMC]

Hi,

1. Make up solutions of the same concentration of each single compound; for each compound one acidic solution (where the substances are stable I guess?) and a pH 11 solution. Then quantify the acidic samples. Over a period of time reinject the pH 11 samples.
2. I would not worry too much about the buffering capacity of ammonia as long as the separation works (just as FA at acid pH). It will of course always depend upon the sample size in how far the pH changes. You could probably try that with a pH meter if you had an appropriate electrode suitable for small volumes.
3. Manufacturer of the system?

[unmgvar]

Hi Mattias

I will only add that you do need to check if your HPLC can work at the given PH. the problem generaly resides in the PEEK part, like pump seals and rotor seals, injector ports and such.
over time they could degrade and affect sytem wide, column performance as well as your application.

[Yury Zelechonok]

To Mattias

First, you mobile phase will be very problematic for MS detection with sulfuric acid used. Second stability of the peptide can be an issue since you have no way to know the nature of your particular peak. Is it intact peptide or degradation product and to what extend it was converted (without MS it is difficult to establish)? Third, your instrument components indeed need to be properly selected (see manual for your HPLC system). Some components for example detector flow cell is not recommended to use above pH 11 at all. Having pH 11, you are working on a top of allowable limit and long term performance of the cell can be questionable. Considering all these I would not use this method as a long term analytical solution.
In your question you indicate that pH is powerful tool in obtaining desirable selectivity. When you change pH you are changing charge state and as result hydrophobic properties of the peptides. In acidic conditions certain molecules are co-eluted since they have similar hydrophobic at this particular pH. To be able to resolve such critical pairs without changing pH significantly you need at least one more interaction force. This can be an electrostatic interaction. To obtain this secondary interaction you need a column with well defined ion-exchange properties along with hydrophobic. Promix MP column from SIELC is a combination of two interactions on a single stationary phase. See graphical explanation of enhanced selectivity of mixed mode columns vs. RP columns following this link http://nexep.com/Technology_2D_Properties.html
This column is capable to resolve complex peptide mixtures when components are different in ether hydrophobicity, or ionization constant or both http://nexep.com/Technology_Peptides_Selectivity.html
The typical mobile phases are ammonium format/acetate, TFA for MS or sulfuric/phosphoric acid if you need low UV detection. The MP is always acidic so no need to worry about degradation.

[Mark Tracy]

Most of the components are reasonably resistant to pH 11 for most modern HPLC systems. Quartz, sapphire, stainless steel, PEEK, PTFE, zirconia and gold are typical wetted materials that tolerate those conditions.

But the rotor seal on Rheodyne valves is Vespel by default. Vespel will dissolve slowly at pH 11 and the resulting oligomers will contaminate everything downstream. You can substitute PEEK or Tefzel rotor seals.

Some older systems have Vespel parts, in particular the HP1090 pump and HP1050 autosampler. I believe that more modern replacement parts are available.

Also there is one brand (Shimadzu? correct me if I mis-remember) that uses polyacetal rotor seals in their autosampler, and that too is vulnerable to high pH.

I have also seen one case where pump seals would bleed MS-detectable stuff at high pH. I don't know if that was just a bad lot or if it was bad design.

[Mattias]

Thanks for your replies!

I work with an Agilent 1100 system, and I havn't modified anything yet. The peptides are not exposed to the high pH until they are injected. I guess an on-column degradation will show in split/broadend peaks??

I have tested C18, phenyl and nitrile columns from different suppliers, different solvents, different buffers and temperatures. I have used Dry-Lab to evaluate my testruns.

The high pH was the only thing that worked, since it moved the three acids to the front of the chromatogram, making the area around the main peak less crowded. The problem I see now is that the peak shape of the acids are quite poor (tailing > 2.0), the other peaks look fine.

This Primesep column is new to me, but it sounds interesting. There seems to be a supplier in Sweden, so maybe I will try it out if my approach fails.

[Uwe Neue]

We routinely work at high pH, and I have worked with roughly your ammonia concentrations (about 100 mM) in preparative chromatography. No, the HPLC system does not dissolve. Silica columns dissolve, but nothing else is a problem. I know that your XTerra columns are doing fine, and I believe that the Zorbax Extend will do fine as well.

We have done peptide separations at alkaline pH. Not unexpectedly, you will get large differences in the elution patterns. Switching the pH from acidic to basic is the most powerfull tool to scramble up your chromatogram, if you are dealing with ionizable compounds. Changing the pH is also an excellent tool to improve loadability in prep.... But I digress...

[Mattias]

What can I do to improve the peak shape of my early eluting acids?? Now I have a tailing of 2-3 and I need to quantify these peaks down to 0.05%...

I guess it is silanol effects that I see at this high pH? I tested to add ion-pairing agents, but that altered my separation totally.

[SMC]

Hi,

You may be right that even at this high pH (11) your peptides might be partially protonated at the N-terminus & @ basic ligands (such as from R or K); in this case electrostatic interaction between peptides and deprotonated silanols could be possible.

How does the tailing look like? Is it a kind of exponential tailing (which could occur at small sample size due to a dual retention mechanism) or is it an overload tailing (right angle triangle peak shape). At high pH acidic groups at the C-terminus & at acidic ligands are mostly deprotonated (i.e. negatively charged). Overloading can occur readily for ionised compounds, even at diminishing small sample sizes. Try different concentrations and look at peak shape as well as changes in retention (one feature of overloading is the reduction of retention with increasing sample size). However, overloading is a function of k and thus I am surprised that your later eluting compounds don’t show this tailing, provided all peptides have the same conc.? Then again the charge on early eluting hydrophilic compounds is probably higher? Have you got the sequence of the peptides?

May I ask how you quantify your peptides?

[Mattias]

Hi!

The tailing is more of the triangular shape (overloading?). I have injected the same amount for all impurities, which is about 0.25 µg (MW about 1000). I use a 3.0 mm i.d. column. I didn't expect to see tailing because of overloading, but maybe it is possible?

I see this tailing behaviour in my standard Dry-Lab gradients (e.g. 5 - 90 % MeOH in 60 minutes. All later eluting peaks have a tailing of 1.0-1.1. I haven't run this analysis in isocratic mode, but I guess the tailing would be enormous. I will test to lower the injected amount and see if the peak shape improves.

I am afraid I cannot reveil the sequence of the peptide but it does not contain Lysine or Arginine.

The quantitation of the impurities are done via UV-area and external standard curves.

[Uwe Neue]

If you have tailing with early eluting peaks, and no problem with the late eluters, it could be caused by a mismatch between your sample solvent and the initial mobile phase composition of the gradient. This could either be an organic solvent in your sample, or it could be the pH of your sample.

[Mattias]

If you have tailing with early eluting peaks, and no problem with the late eluters, it could be caused by a mismatch between your sample solvent and the initial mobile phase composition of the gradient. This could either be an organic solvent in your sample, or it could be the pH of your sample.

That is probably exact what it is! I didn't think of this...
The sample is dissolved in a pH 4 buffer and injected into a pH 11 mobile phase. If some of the molecules don´t get deprotonated immediately the will elute later and probably cause the tailing.

Now is the question how I can solve this? It is impossible for me to dissolve the sample in pH 11 due to degradation of the sample.

[Uwe Neue]

First I suggest that you prove that this is the problem. You can make a small amount of sample at pH 11 and inject it immediately. If the problem goes away, it is the sample pH.

If this is the case, this can potentially be resolved by starting your gradient with a still higher ammonia concentration, and then reducing it for the actual separation, i.e. the gradient run itself. It depends on your system if you can execute something like this. We have done games like this for preparative chromatography.