Variable Peptide Recoveries

[Eric Moore]

Hi -

I have a typical peptide RP method - ACN/H20/TFA, gradient, C18 column - and I'm getting good chromatograms. However, my recoveries are all over the place. They're very variable (which I'm not too concerned about, since it's a very early-stage method) but what is more troubling is the trend I see based on the diluent I'm using.

I've already ruled out pump/injector/instrument errors.

When I prepare solutions in WATER, the recovery is much higher. When I prepare solutions in MOBILE PHASE (AcN/H20/TFA) the recovery is lower. What I mean is, a sample in WATER might have an AUC of 370. The same concentration in MOBILE PHASE might have an AUC of 220.

I'm using polypropylene HPLC vials. I did an experiment where I tried blocking the polypropylene with BSA, but this showed that the vials - given samples in the aqueous diluent - are not alone in causing this problem.

Could the acidic diluent be causing my peptide to adsorb to the polypropylene (perhaps by protonating it and making it more hydrophobic??)

The peptide is TFA salt. Could the TFA in the diluent have some negative effect on the peptide?

I can run more experiments to empirically figure out the best course to take, but I'd be interested in the theory behind what's happening.

Thanks, all.

Eric

[danko]

Hi Eric,

How much acetonitrile is there in the solvent, in the case of dissolving the sample in mobile phase?
Also, does the injection peak get larger when the sample is dissolved in mobile phase?
Finally, is the described tendency reproducible?

Best Regards

[Eric Moore]

Hi, Danko -

There is 98% water, 2% acetonitrile (0.1% TFA) in the diluent.

I'll have to check on the injection peak.

The tendency is reproducible - sample in water has ~50 - 70% higher peak area than sample in above diluent.

Thanks!
Eric

[Uwe Neue]

interesting... could be the same phenomenon as a wetting/dewetting problem on a C18. See if the MeCN or the TFA or both are the cause!

[Eric Moore]

Is there a link to the wetting/dewetting phenomenon?

Eric

[Uwe Neue]

It has been demonstrated and explored on porous C18 packings. See the following publication:
Mechanism of retention loss when C8 and C18 HPLC columns are used with highly aqueous mobile phases
Journal of Chromatography A, Volume 1075, Issues 1-2, 20 May 2005, Pages 177-183
Thomas H. Walter, Pamela Iraneta, Mark Capparella
However, I can see that something similar can happen on a hydrophobic polymer.

[Eric Moore]

Hi, all -

I found a paper that said sticking of peptides containing high levels of proline and glycine is common in CE (fused silica capillary). These investigators solved their problems by controlling the pH of their samples.

I'm not very familiar with CE. However, about 40% of the amino acids in my peptide are either proline or glycine. Could I be seeing the same phenomenon here? Do fused silica capillaries and polypropylene HPLC vials share similar hydrophobicity, or other characteristics, that would be causing this?

Eric

[danko]

Hi Eric,

I’ve heard and keep hearing stories about sticking protein/s to vials thousands of times. 99.9 % of the instances there is nothing to it. The beauty of this story though is the simplicity – it’s easy for the brain to adsorb Aren’t we talking about 30 to 50 % loss?!
Anyway, if you want to investigate the matter, you can do it as follows:
1. Prepare some/several milliliters solution of your peptide in a mobile phase.
2. Fill 2 HPLC vials with this solution.
3. After some minutes (20 – 30?), empty the content of one of the vials and then fill it up again with the original peptide solution.
4. Analyze both vials and compare the resultant areas.

If there is some adsorption going on the vial that was emptied and then refilled should be the one that results in the largest area – because the vial’s inner wall was saturated with the peptide and couldn’t possibly adsorb as much peptide molecules as the vial that wasn’t saturated.

I’m inclined to expect the same area plus/minus the natural variation.

As I mentioned earlier – although indirectly - the chance is much higher that the peptide is degrading due to the low pH – acid induced degradation. And if the degradation products are highly ionized there is this possibility of them eluting with the void volume.
So, it might be a good idea to check the injection peak for any changes/increases roughly proportional to the main peak decrease.

Best Regards

[Eric Moore]

Thanks, Danko, I'll give that a try.

As for low pH degradation, some data indicates it isn't the culprit. There is no trend with regard to peaks in the void volume. However, I did check the pH of some of my samples after a run (an experiment looking at peptide stability in base.) I'd diluted these samples in Mobile Phase A (pH = 2) before loading them on the injector. The samples with the lowest pH that sat the longest on the sample tray before injection (therefore in the acidic diluent the longest) had higher peak areas than the low pH samples that were injected earlier in the run!

If degradation were happening, I'd expect to see lower peak areas for these samples (along with degradation peaks.)

It's all very confusing ... the data aren't showing any obvious trends, except for that if I dilute samples in water as opposed to acidic mobile phase, I get higher values for the same concentrations. And at higher pHs (9 - 10) I get higher values and lesser variability than at acidic pHs. Which DOES seem to indicate degradation at low pHs.

All the data are extremely variable. It almost seems like an injector problem, but I've already checked that twice! (Weighed a vial of water. Injected multiple injections [20] then checked the average injection volume. It was dead on.)

Very weird.

Eric

[tom jupille]

If you have more than one peak of comparable size in a chromatogram, injection variability is easy to prove/disprove. Take the ratio of two peaks and compare the variability of that ratio to the variability of the individual peak areas. If the areas look *better* than the ratios, then you don't have an injection volume problem (that kind of error would cancel out and make the ratios look better).

[Eric Moore]

Unsatisfying Conclusion:

Basically, when I use water to dilute everything, and avoid acids or bases, my reproducibility is a-OK. This peptide is only a model for us, and so The Powers That Be deemed it not worthy of me spending more time on figuring out "why."

Degradation? pH-dependant adsorption? Small elf or gnome residing in the HPLC? Your guess is as good as mine ...

Thanks for all the input. This is a TREMENDOUS resource for me.

Eric