HILIC separation for sialylated glycans

[Moe89]

Hello guys,

First of all, thanks for the help the last time!
I'm having trouble with the separation of sialylated glycans in HILIC mode. The aim is to analyze the glycan structure of gylcoproteins.
The run for the "Glucose Homopolymer" standard works fine, giving me GU values for the detected peaks.
I run the same setup on the "Sialylated biantennary library" to look particulary for sialylated glycans, but I can't separate the peaks.

Column:
Luna NH2 100A (Phenomenex), 250x2mm

Buffer A: 100% ACN
Buffer B: premixed ACN with ammonium formate (50mM, pH 4.4) 50:50
Buffer C: same as B, but 90% ammonium formate, 10% ACN
Temperature: 30°C
Injection vol: 5µL
Standard "Sialylated Biantennary Library" is dissolved in starting condition (60% ACN, 40% ammonium formate)

Gradient:
Flowrate is always 0.5 mL/min. Stoptime is 65mins.
0mins 60%A 40%B
40mins 0%A 100%B
50mins 0%A 100%B
52mins 0%A 0%B 100% C
58mins 0%A 0%B 100% C
60mins 60%A 40%B
65mins 60%A 40%B

I attached the chromatograms for both standards.

[Andy Alpert]

First of all, did you attach anything to the reducing end of your glycans? I have in mind something like aminopyridine or some other tag that may have a (+) charge, which would cause your analytes to migrate in an oriented manner through the column (sialylated end down, reducing end up). That would make your chromatography more sensitive to the presence or absence of a sialic acid residue.

It looks like you're getting partial separation of the variants in each cluster of glycans. How do these differ from each other? If it's by the number of sialic acid residues, then use less salt in your mobile phases in order to tune up electrostatic attraction. Currently you're running a gradient from 10-25 mM ammonium formate at the same time that you're running a decreasing gradient of ACN. Try keeping it at 10 or 15 mM throughout the run. I'd suggest that you have the starting mobile phase (now 60% A/40% B) premixed in your reservoir bottle instead of generating it by blending online with 100% ACN. Same goes for the final mobile phase. That permits you to control the concentration of the salt.

if the nearby peaks differ by an uncharged residue - say, a mannose or GlcNAc residue - then try a neutral HILIC column instead of an amino column. These too will be sensitive to the number of sialic acid residues, although to a lesser extent than will an amino column.

[Moe89]

Hello Mr Alpert,
2-AB is attached to the reducing end of the glycans and is then detected with FLD.
Indeed, we have three forms of sialylation in the standard. There are "asialo" (unsialylated), mono-sialylated and di-sialylated glycans , which could explain the three clusters.

I will prepare freshly:
Buffer A: (80% ACN and 20% ammonium formate, 50mM, pH 4.4)
Buffer B: (50% ACN and 50% ammonium formate, 20mM, pH 4.4)
and set the gradient to 100% A to 100%B in 40 mins. So same gradient as before, but with premixed buffers and same salt concentration throughout the run.

I already tried to decrease the slope of the gradient by extending the time from 40 minutes to 60 minutes, but the peaks just shifted.

Do you think it is worth to vary the pH a little? Or changing the gradient to even higher percentages of aqueous solvent?

Best Regards,
Moe

[Andy Alpert]

Adjusting the pH isn't going to help much. The pKa of a sialic acid is around 2.6, so you'd have to go to a really low pH in order to uncharge them. It isn't necessary anyway, since you could tune down their electrostatic attraction to the stationary phase by including more salt in the mobile phase. If instead you increased the pH, then the only effect would be that the charge density of the amino material would decrease substantially, decreasing the degree of hydrophilic interaction. That is the opposite of what you want. Leave the pH alone.

Running the gradient to a higher percentage of water: This would be helpful only if some of your analytes were not eluted within your gradient or at least within a reasonable running time. Here your last cluster of peaks is eluting right at the end of the gradient, so increasing the percentage of water that you get to is unlikely to benefit the separation.

Your comments tend to confirm the speculation that the peaks within each cluster are glycans that vary in the number of uncharged residues. If the experiment with the salt concentration does not improve the separation, then there are two changes that could help:

1) Switch from an amino column to a neutral HILIC column;
2) If you're using a 5-µm particle diameter column, then switch to a 3- or even a 1.7-µm column. That will improve the separation by increasing efficiency (= peak sharpness) rather than selectivity (= the separation of the peak maxima).

[Moe89]

Thank you Mr Alpert,

your text was very informative and gave a me a clearer understanding. I prepared the buffers yesterday as you wrote in your previous comment and today I will do a run with a constant salt concentration at 10mM and look at the results. I will give you feedback if the run improved or not.

Thank you very much.

Best Regards,
Moe

[Moe89]

Hello Mr Alpert,

The run did not improve with the controlled 10mM salt concentration. The chromatogram was similar to the first one with the clustered peaks, just with retention time shifts.
I will look today if the column is still functional since it is a little bit older and look for new columns for the glycan separation. Do you have any suggestions for a new column ? I though about an amide gel, but maybe there are better functional groups for glycan/sialylated glycan analysis (generally glycoprofiling of glycoproteins).

Best Regards,
Ömer Alkan

[valentina_waters1]

New column for glycans
you could try the waters Glycoprotein BEH Amide 300Å 1.7 µm
very good HILIC column with high resolution and very stable, frequently used for glycopeptide mapping

[Moe89]

Hello,

Do you know if there is a possibility to test the columns before purchase?

Best regards,
Ömer Alkan

[valentina_waters1]

[quote="Moe89"]Hello,

Do you know if there is a possibility to test the columns before purchase?

Best regards,
Ömer Alkan[/quot
Sorry for my delayed response,
this may be possible, please ask for a visit from a waters sales rep . Please send your request here:

http://www.waters.com/waters/en_US/Cont ... cale=en_US

and just let me know if you need technical help