Negative peak in ion exchange

[Mattias]

Hi,

I am testing a weak cation-exchange column using an ammonium acetate buffer gradient and I get a very sharp and nice peak of my peptide. The only problem is that just before my peak, there is a large negative peak (shark-fin styled) that disturbs the integration. And this is no way near the front of the chromatogram (k' about 10)

So something else in my sample is retained. But my sample is just the peptide as acetate salt dissolved in 50% acetonitrile.

Any ideas what it can be?

[Andy Alpert]

What are your mobile phases and what wavelength are you monitoring? Also, if you're using an automatic sample injector, then what are you using as a needle wash solvent?

[Mattias]

Hi Andy,

I simplified my system in the first post, since I have coupled a WCX-precolumn together with a HILIC-column.

The reason for the WCX-column is to clean up my peptide before the HILIC (but so far I have only used pure standard solutions).

The following setup has been used: The sample is loaded on the WCX-column (2.1x20 mm) with a mobile phase of 50% acetonitrile in water. Then the flow is switched and the precolumn is washed backwards with the initial mobile phase of the HILIC system (20 mM ammonium acetate pH 4.0 in 90% acetonitrile). Then I start a gradient to end at 30 mM ammonium acetate pH 4.0 in 75% acetonitrile.

I have done the same separation with a RP-precolumn before, and there I do no see any negative peak. The problem is that the unwanted proteins also stick to the RP-column, and they do not on the WCX-column.

I will change my HILIC-gradient and add an isocratic step in the beginning to see if the negative peak can be moved away just a minute or two. Then I have a system that works. The system is a Agilent 1260 without any needle wash and I use 240 nm.

When I have you on the line so to speak: I am writing an article about HILIC separations of basic peptides. Do you know the reason why there is no tailing of these peptides in HILIC? Compared to reversed-phase, where we see a lot of issues with peak shape. If you have a reference I can use, that would be great!

[Andy Alpert]

This feels like playing 20 Questions; every response provides a little more information that moves us closer to the answer...

It seems that your peptides are particularly basic. That leaves open the possibility that they have acquired some anions from the mobile phase to serve as counterions. In this case that would be acetate. You didn't specify if the peptide peak and the negative peak were observed eluting directly off the WCX column/cartridge or if you observed them in the subsequent HILIC run. If it's the HILIC run, then the peptide may have depleted some acetate from the mobile phase and then eluted more slowly than the portion of the mobile phase now lacking in some acetate ion. That would result in a negative peak sometime before your peptide peak, assuming you're monitoring absorbance at a wavelength that's absorbed by acetate (< 240 nm). You didn't tell us what wavelength you are monitoring.

Another possibility: Since your WCX mobile phase has no salt added, then when a sample of this mobile phase (in this case, the fraction containing the peptide that just eluted from the WCX column) reaches a HILIC column that's being eluted with 20-30 mM acetate, then the slug of liquid from the WCX column will have less acetate in it than the surrounding mobile phase and will have lower absorbance, at least if you're monitoring a wavelength that acetate absorbs. That accounts for your negative peak. Do you still see it in a blank run (i.e., with no peptide)?

Tailing of basic peptides in HILIC or lack thereof: Tailing of these in RP has usually been ascribed to interaction with silanols. Traditionally this has been handled by including an amine in the mobile phase to saturate the silanols and shield them from the basic groups in peptides. As for HILIC: I assure you that basic compounds can elute in badly-shaped peaks there as well. It depends on how much salt you have in the mobile phase. For moderately basic compounds, the 20-30 mM salt that you're using suffices. For compounds with a higher charge-to-mass ratio, such as arginine, you should use 40+ mM salt. For an extreme case such as aminoglycoside antibiotics (kanamycin, etc.), then you need over 100 mM salt in order to get symmetrical peaks in HILIC. For a demonstration of this effect, see Fig. 14 in my 2008 paper that introduced ERLIC. The problem involves a situation where molecules of the same compound have more than one possible counterion. The resulting ion pairs differ in polarity (or, more precisely, their degree of hydration) and migrate through the column at different speeds. Result: Skewed peaks or even multiple peaks for a single analyte. Using a high concentration of salt in the sample solvent and the mobile phase insures that all of the molecules of the same type will share the same counterions. Don't blithely desalt a peptide on a C-18 cartridge using TFA and then dry it down and shoot it onto a HILIC column when the mobile phase contains acetate or formate.

[Mattias]

I agree and apologize that the information about the system in the first post was very poor..

It is likely so that the buffer free mobile phase in the trap column is causing the negative peak. I see the same peak in a blank injection (detection at 240 nm). I am a bit surprised to see that the "buffer-free plug" is moving that slowly through the column. I have changed the gradient a bit, and now I have a resolution of about 5 between my peptide and the disturbance, so everything is OK now.

This particular peptide is probably the most basic that I have worked with. It contains an agmatine-group with a pka of about 12-13. And with this peptide I get an asymmetry factor of about 1.5 in HILIC.

The "normal" basic peptides that I work with contains lysine and/or arginine. And with this HILIC system (20 mM ammonium acetate) I usually get assymetry factors of 1.0 - 1.1. If I inject the same peptides on reversed phase, the assymetry factors can be 4-6.

So HILIC is a very good option for these peptides. I assume that this is due to that the peptides mostly interact with the water layer, and not the column itself? The HILIC column that I use is pure silica (and thus only silanols).

It is very interesting what you wrote about counter-ions. I have seen this phenomenon before when I started with HILIC and only used 5 mM buffer in the mobile phase. Maybe I should start to add 20 mM of buffer in the samples as well, have not done that so far.