Question about HILIC-MS for phosphopeptides

[squall]

Hi All, I am recently doing some HILIC-MS for phosphopeptides separation and identification. Ideally, HILIC should retain better than C18 for phosphopeptides, but compared to the C18-MS data, I really don't have expecting results from HILIC.

Maybe I didn't get the optimum condition for HILIC, or using HILIC itself just can't have a very good result for LCMS phosphopeptide identification?

My purpose is to obtain some data showing that using HILIC-MS I can have some more information which is not acquired by (C18)LCMS, such as some multiply-phosphopeptide separation, or a better separation between non-phospho pep and phosphopep. Based on current data, I am kind of not sure whether this goal is achievable or not?

I tried to Google some reference, but not exactly find what I want. Does anybody see some references (either paper or application notes) really showing that using HILIC LCMS can have a better or different story than regular C18 LCMS regarding to phosphopeptides or phosphoproteomics?

Thanks

[Andy Alpert]

No single method works well for identification of phosphopeptides from crude digests. There are two reasons:
1) Phosphopeptides tend to be of lower abundance than unmodified peptides and so tend not to be above the cutoff threshold for any single MS scan window.
2) Phosphopeptides tend not to ionize as well as unmodified peptides, again not being identified easily as a result.

The conclusion from this is that anything you do to separate phosphopeptides from unmodified peptides upstream from the MS is going to increase success at their identification greatly. That means 2 dimensions of chromatography, whether you use RPC or HILIC as the final dimension. Since Steve Gygi's work in 2004, the most common sequence has been SCX-RPC. Adding an affinity isolation step for the phosphopeptides in the SCX fractions, whether titania or IMAC, increases id's about 3x more than what you find without it. Recently SCX has been challenged as the upstream step. A sequence of high pH-low pH doesn't have any particular affinity for phosphopeptides but spreads out all peptides well enough that it increases the total number of id's for both phosphopeptides and unmodified peptides. High pH anion-exchange has also been investigated. Perhaps the most promising combination is an anion-exchange material being used in the HILIC mode, a combination called ERLIC (Electrostatic Repulsion Hydrophilic Interaction Chromatography). Using the right conditions, this combination does have affinity for both phosphopeptides and sialylated glycopeptides, pulling them away from the bulk of the unmodified peptides. Volatile mobile phases can be used. The results with phosphopeptides are decidedly better than with plain HILIC. A paper compares SCX-RPC, HILIC-RPC, and two varieties of ERLIC-RPC: X. Chen et al., J. Chromatogr. B, 879 (2011) 25. The HILIC-RPC combination was the worst for phosphopeptide identification.

ERLIC works quite well for multiphosphorylated peptides, although eluting any with > 3 phosphates or many of the ones with 3 phosphates may require nonvolatile salts such as phosphate. See my poster from HPLC 2008 on the subject, per the following link: http://www.polylc.com/downloads/HPLC_20 ... slides.pdf
Siu Kwan Sze and colleagues have published a number of papers since 2008 on the use of ERLIC with phosphopeptides. One that's currently in peer review involves the use of ERLIC capillaries as the final mode before MS, per your speculation [it does work]. Also, Amanual Kehasse had a poster at ASMS 2011 (MP 484) [available from me upon request] demonstrating about 60% more phosphopeptide id's using an ERLIC-IMAC-RPC sequence than with an SCX-IMAC-RPC sequence.

[squall]

Thanks a lot, Andy. From reading, I believe that the combination of different columns, especially the ERLIC can increase the phosphopeptide identification. But for my case, I already did the enrichment, which means I am able to directly identify a certain number of phosphopeptide using C18 LC-MS/MS. Based on this, if I Just switch the C18 to HILIC (any kind of HILIC), and do LC-MS/MS. nothing else is changed (of course the gradient need change), and no LC combinations, do you think the HILIC one could have a better results than C18 one either on "multiply-phosphopeptides" or "more separation between phospho-pep and non-phospho pep?"

Thanks a lot

[Andy Alpert]

Not necessarily. Certainly phosphate groups are hydrophilic and promote retention of peptides in the HILIC mode. The trouble is, the disparity in polarity isn't all that great. I'm not impressed by the enrichment factor that McNulty and Annan reported for HILIC of phosphopeptides, so they won't be separated as a class from the unphosphorylated peptides. As for multiphosphorylated peptides: Sure, they'll be well-retained in HILIC. Chances are they'll be well-retained in RPC too. Multiphosphorylated peptides tend to have one or two missed cleavages (the phosphate group tends to inhibit tryptic attack nearby), meaning that they're rather large for tryptic peptides. That means that they'll have a lot of residues of all kinds, including hydrophobic ones.

So: You enriched first for phosphopeptides? Sounds like you're following Huilin's example from 2008. The trouble with this is that peptides with > 2 phosphates tend to get lost during enrichment on either TiO2 or IMAC. That 2011 study by Chen et al. that I cited above didn't find a lot of them. They had pre-enriched as well, and speculate that they lost the multiphosphorylated peptides at that stage. Maybe that's why you haven't been finding them using a C-18 capillary; they aren't in your sample anymore.

Why not use ERLIC? You could pre-enrich with it , collecting several fractions (which is not feasible with high-affinity methods like TiO2 or IMAC), including the multiphosphorylated peptides, and then run each one through a C-18 capillary. That will break up the large set of phosphopeptides into subsets of a more manageable scale, with fewer peptides in each scan window. ERLIC does a nice job of separating the phosphopeptides from the unmodified ones, so if you're determined to use a HILIC capillary in place of C-18, then consider making it an ERLIC capillary.