Regenerating polyhydroxyethyl columns

[DJ]

I'm getting poor peak shape with one of these columns. If flushing with 50 mM HCOOH O/N does not correct this, is it likely the column has developed a void?

[tom jupille]

If your chromatogram has more than one peak and all peaks show the same problem, then a void space is a probable cause. If different peqks sjhow different problems, then a chemical problem iw more likely.

[Andy Alpert]

DJ -

1) What is your starting mobile phase and what solvent is your sample in?

2) If this has something to do with those peptides you mentioning earlier with 9-10 Arg residues, then you're not using enough salt in the mobile phase. It's also necessary to match the salt in the sample solvent with the salt in the mobile phase. Read my 2008 paper on ERLIC for a discussion about this, per the following link: http://pubs.acs.org/doi/pdf/10.1021/ac070997p

In brief: In order to get symmetrical peaks for polyelectrolytes in HILIC, all of the ionizable groups must have the same counterions. Different counterion combinations will differ in polarity and hence will elute at different times, leading to tailing peaks or even multiple peaks for a single solute. For something as highly charged as an aminoglycoside antibiotic (e.g., kanamycin), you must have at least 120 mM salt (ammonium acetate or some such) in the mobile phases and in the sample solvent to get a symmetrical peak.

Andy Alpert

[DJ]

Thanks, Andy. I'm very familiar that Anal Chem paper

Your explanation for TEAP vs methylphosphonic acid differentially affecting retention of basic peptides makes sense to me, but salt type, content in HILIC/ERLIC is a mystery to me. Some swear at least 10 mM salt is necessary in hilic/erlic, but I actually had good luck using WAX in ERLIC mode in a gradient from 90% MeCN/0.1% HOAc to 50% MeCN/0.1% HCOOH.

For the polyhydroxyethyl column, I used 20 mM TEAP in the mobile phase. I start at 90% MeCN. The separation looks great on a second (nearly identical) HILIC column.

I have heard recovery from HILIC is excellent to quantitative. I have used a column for analysis of synthetic peptides that are similar to ones I am now I am trying to isolate from a natural source. Contamination is a huge concern. I wouldn't even think about using a RP column that has previously seen synthetic peptide. But what about HILIC? Do you suppose I could get rid of any last traces of synthetic sample by flushing with high salt or 50 mM HCOOH o/n?

Same deal with WAX?

[Andy Alpert]

Loss of sample is usually attributed to some combination of the following:

1) Adsorption to the metal components of a column or HPLC system. This would be a particular concern with peptides like yours with multiple charged residues [NOTE: What kind of peptide has 9-10 Arg residues? A fragment of an ion channel lining protein?] This can often be reduced or eliminated by overnight treatment of the column and HPLC system with 40 mM EDTA.2Na. An alternative is a couple of shots of albumin followed by a salt gradient. While that is quite effective at reducing "hot" adsorption sites, you may find traces of albumin popping up in your mass spectra in collected fractions.

2) Precipitation of components that are marginally soluble in the starting mobile phase. At 90% ACN, you're right on the edge with some peptides.

3) Hydrophobic interaction. While high organic solvent levels reduce this to a negligeable amount, you could conceivably run into trouble if a peptide had appreciable hydrophilic and hydrophobic domains and remained on the column during a gradient to low % ACN. That's why we take particular care to get rid of any hydrophobic character in a stationary phase that's to be used for HILIC. PolyWAX LP is as hydrophilic as any anion-exchange material in the industry, so go ahead with your project.

[DJ]

Loss of sample is usually attributed to some combination of the following:

1) Adsorption to the metal components of a column or HPLC system. This would be a particular concern with peptides like yours with multiple charged residues [NOTE: What kind of peptide has 9-10 Arg residues? A fragment of an ion channel lining protein?]

What kind of peptide? A peptide that has been a real devil to synthesize and purify! Has 4 mets, and an inordinate desire to undergo sulfoxidation (which complicates natural ligand isolation). It has 10 Args, but pI is between 7-8. It is loaded with hydrophilic residues (D, E, N, Q), yet, its retention time in RP-HPLC is close to BSA. Among a pool of 60 or so neuro peptide sequences, the sum of HILIC retention coefficients predict it would be one of the best retained peptides in this mode as well--so, binds RP, cation exchange, and HILIC columns fairly well. Also never met a wetted surface it didn't like. For all these reasons, I'm looking for methods (SCX, HLIC) which minimize sample loss.

This can often be reduced or eliminated by overnight treatment of the column and HPLC system with 40 mM EDTA.2Na. An alternative is a couple of shots of albumin followed by a salt gradient. While that is quite effective at reducing "hot" adsorption sites, you may find traces of albumin popping up in your mass spectra in collected fractions.

Most of the pump has titanium parts. I use PEEK pre-column frits, tubing, and frequently change the pre-column frit. I'm less worried about loss of sample in hilic to column frits (these have seem multiple injections of peptides.).

2) Precipitation of components that are marginally soluble in the starting mobile phase. At 90% ACN, you're right on the edge with some peptides.

I use 90/10 MeCN/HOAc to extract this peptide from tissue samples. The high MeCN is very exclusive compared to aqueous or acidic methanol exctraction. No solubility problems at 90% MeCN, but could easily reduce this to 80%, as this peptide elutes below 60% or so MeCN in HEA.

3) Hydrophobic interaction. While high organic solvent levels reduce this to a negligeable amount, you could conceivably run into trouble if a peptide had appreciable hydrophilic and hydrophobic domains and remained on the column during a gradient to low % ACN. That's why we take particular care to get rid of any hydrophobic character in a stationary phase that's to be used for HILIC. PolyWAX LP is as hydrophilic as any anion-exchange material in the industry, so go ahead with your project.

Interesting. I never considered material may stay bound via hydrophobic interaction. Would you ever consider "cleaning" these hydrophobic components in the same way you would a RP column-- 100% IPOH, then, perhaps DCM, even THF?

[Andy Alpert]

A reversed-phase column will interact with the hydrophobic residues of your peptide and will ignore the hydrophilic ones. A HILIC column has the opposite selectivity; they "see" different parts of the molecule.

I wouldn't worry about cleaning up a HILIC column in the manner you suggest. The high starting level of organic solvent should take care of that. Now, if you try using some of the "tunable" stationary phases that are being promoted for HILIC and which contain appreciable hydrophobic character as well as hydrophilic character, then I would worry about this.

Just because your HPLC system has a lot of titanium components doesn't mean that they won't interact with proteins or susceptible peptides. Titanium does resist attack by some reagents that corrode stainless steel, but that doesn't mean it's completely inert. Pure titanium is too brittle for most practical applications. The commercial metal is an alloy, frequently containing several % aluminum. It will adsorb stuff. This is what passivation with EDTA prevents, not attack by chloride ion on the titanium.

Concerning your sticky peptide: sounds like Lung Surfactant Protein. Its role in nature is to stick to the alveoli in your lungs, and it sticks to everything else as well. The only chromatography method that worked was a gradient of NaClO4 (a chaotropic salt) in 70% ACN on a PolySULFOETHYL A column at pH 3.