Preparative IP-RPLC of underivatized amino acids

[James]

I want to develop a preparative IP-RPLC separation of the more polar proteogenic amino acids (underivatized), particularly gly, ala, pro, hyp, glu and asp. The IP reagent must be removable, ie volatile etc.

I have been developing an initial method along the lines of Kostas Petritis’ analytical method (Jr Chrom A, 833 (1999) 147-155) (thank you Kostas!) using PDFOA as IP reagent and trying a number of prep scale C18 columns: Atlantis, Symmetry, Polaris and Supelco HS F5. However, I am finding I do not achieve baseline separation of asp/hyp and ser/glu and gly, ala & pro are all very broad peaks (4-5 mins each!).

I would welcome any suggestions on improving the method to get better resolution. Can anyone suggest prep-scale columns that may provide better retention/chromatography and/or alternative IP reagents suitable for underiviatized amino acid isocratic separations?

[SIELC_Tech]

As alternative you might look at Primesep. There is no need for IP reagent and you can use ACN-water-TFA as your mobile phase (although somebody might argue that TFA is an IP reagent). IP regagent is attached to the surface of silica. The methods below has some of your acids.
What is objective of your experiments? Are this exact structures of your AA or you have derivatives/ analogs?

http://hplcmethods.com/compound_002.php

http://hplcmethods.com/application_087.php

Sulfuric acid in the second method can be replaced with TFA-this method was developed for on eof our customers who needed UV method. 0.1% of sulfuric acid is equivalent to 0.15% of TFA

Regards,

Vlad

[James]

Thanks for your reply Vlad; my aim is preparative separation of amino acids for subsequent radiocarbon dating and offline stable isotope analysis from archaeological material. I am working with acid hydrolysed amino acids from ancient proteins hence I cannot use derivatization and need to remove all IP reagents or other extraneous carbon containing components after separation.

Primesep looks an intriguing possibility and I will get hold of an analytical column to test. Do you think scaling up will cause any problems? I should mention I am using UV detection not ELSD, I have a fraction collector and ideally want to collect about 2mg of each aa.

I am mainly interested in gly, pro, hyp, arg, ala, glu, asp, lys and ser or as many of these as poss. Two separations is not a problem if necessary as accuracy, rather than time, is my main concern. I have been wondering whether there may be any way of derivatizing amino acids with non-carbon based additions but have no experience with derivatization and must admit cannot think of anything that might work.

Many thanks, James

[SIELC_Tech]

James,

If you are looking at UV detection then TFA might be a problem-you need to use very clean TFA as UV activity of amino acids is not that great. What else you can tolerate in the mobile phase (sulfuric acid or buffer, phosphoric acid of buffer). Going to more transparent buffer will help your detection.

For prep separation you might need 5 um material too if resolution between acids is not good enough. You can try to do small scale prep work on analytical column (4.6x250 mm). I will try to see if we have a method for these amino acids (or can run the in the lab). Whatever you see in our method you will be able to reproduce, if you add more amino acids in the mix then you might have co-elution of some and need to play with conditions. Can you derivatize your acids (esterify or put BOC for example), separate acids by prep chromatography and then clip off protection?

Here is the link for prep article using Primesep (not amino acids):

http://www.sielc.com/Literature_Posters.html

Regards,

Vlad

[James]

Vlad, yes mineral acids and the inorganic buffers you mention would be fine for mobile phase so i'll play about with them. Just talked with Hichrom, your UK distributers for Primsep columns and they are going to let me try out an analytical column so will have a play. Any suggestions for alternative mobile phase conditions for UV detection would be great.

I have been trying to avoid derivatization for a number of reasons primarily because, for stable isotope analysis, there is the potential problem of fractionation; ie altering the distribution of carbon isotopes whihc is one of the things we measure. The derivatization and clipping off of derivatization agents would need to go to 100% completion to avoid this and I would need to completely remove all carbon containing derivatizing agent afterwards. I havent looked to far into this but my initial feeling is getting these reactions going to 100% completeness, and comlete removal afterwards, would be difficult.

[SIELC_Tech]

James,

Your amino acids will retain/ separate based on hydrophobicity (almost none for a lot of acids) and ion-exchange interaction. For UV detection you can use ACN-water with sulfuric acid, phosphoric acid, corresponding Na, K buffers. IN addition to that you can play with pH of the mobile phase, this will slightly change ionization state of the stationary phase and your amino acids and might provide you additional room for separation. Nature and concentration of the buffer will play major role in your separation. Try first to use isocratic conditions ACN-water-sulphuric acid (10/90/0.1) if your retention is not good enough you can switch to phosphoric acid. For buffers you can try 5-30 mmol at different pH (pH=2-2.5 for example). If you have ELSD somewhere I would try to use it with TFA instead of inorganic buffer, in this case you can develop a method, do prep separation and then just remove your solvent and TFA by simple evaporation. If you use buffers you will need to remove them after prep chromatography (unless you don't care about it)


Vlad

P.S. We are conducting series of seminars in UK at the end of October. Please check with Hichrom, may be they already have one at your company or can set up one at your location.

[Kostas Petritis]

Hi James,

The analytical problem you are describing sounded familiar and after some thinking I found out why. It happens so that you submitted a manuscript to a journal and I was one of the reviewers so I am very familiar with the details of your experiment and what you are trying to do.

Although I have already send my reply to the editors, discussing anything more on the subject (providing more scientific input) might present a conflict of interest to me if ever I will receive a future manuscript from you.

We can continue the discussion off forum, but maybe I can say two or three things staying on publication facts. As you can see from the publication that you mentioned, there are several columns tested and only one (Merck, Purospher RP-e) gave good base line separations of all the polar amino acids. I had tested maybe 20 different C8-C18 columns and none of those gave the same results (some came close but still). If you can not use the stationary phase of the publication (due to high prices or unavailability) maybe you can use the software that Tom mentioned that can find columns with similar characteristics (although you are not guranteed that you will achieve same results under ion-pairing conditions).

Furthermore, have you seen one of my previous publications (Petritis et al. J. Sep. Sci. 2002, 25, 593-600) where I use an analytical column to do an LC-ELSD/NMR experiment. When I inject 150 ppm I got excellent peak shapes and baseline separation of 14 underivatized polar amino acids under isocratic conditions. When I overload the column with 2000 ppm of each amino acids, peak shapes really deteriorates. This implies that if you use a preparative column and stay within the limits of how much you can load you shouldn't have any problems (although you might have to inject several times smaller quantities and collect them...).

Furthermore, there is some work done with mixed columns (look at Chaves-das-Neves and co-workers, J. High Resol. Chromatogr. 20, 1997, 115, he has another article as well in Anal Quim I think...) so Vlad's approach might work as well (it would be nice to have his column when I was working on these projects).

I have also used HILIC approaches (using columns from PolyLC with some success).

Anyway, interesting problem... hope the above helps,

Kostas

[HW Mueller]

What interests me here is why organics bother in your isotope ratio. Are you converting to CO2 and analyzing that? We have failed in getting rid of unrelated organics in a isotope ratio attempt via CO2.

[James]

Underivatized amino acids separations is a small world! I take your point Kostas about the publication, thanks very much for the suggestions and refs however, likewise Vlad for the Primsep info.

Hi HW Mueller,
I am worried about the derivatization process (and subsequent decoupling after separation) causing fractionation and altering the original amino acid carbon isotope ratio. But as you suggest this could also happen with incomplete removal of the reagent. After separation I convert the whole molecule to CO2, via an elemental analyser, as you suggested.

When you say you have had problems getting rid of ‘unrelated organics’ do you mean solvent molecules or derivatizing agents?

James

[HW Mueller]

Yes, the decoupling could be a major problem, unless you separated the amino acids (now a single one) from all "decoupling" products. etc. Even then one can get carbon contamination from solvent, reagents, CO2 from the air or any other carbon around (in other words, I am not sure where the C contamination originated). Of course, it all depends on the accuracy you want.