Problem with gradient HIC

[qrys]

I normally work with isocratic elution of proteins using PEG coated silica columns. The mobile phase is phosphate buffer pH 7 with various salt concentration of 0-1 M Na2SO4 in the phosphate buffer. Thus I have collections of data of a protein (e.g. lysozyme) ran through the column at different salt concentrations.

The following are the peak width of lysozyme that was isocratically eluted with various concentration of Na2SO4. Peak width is measured at the baseline.
1 min (0.2 M)
1.5 min (0.5 M)
3 min (0.7 M)


I then ran a single protein on the same column with gradient elution (HIC). The result was bad as I obtained a peak that was significantly broader than the isocratic peak. Peak width as measured at the baseline was about 4-6 min.

The condition for gradient HIC is the following:
PEG coated silica column, ~ 5 cm height, 5 mm diameter
mobile phase: phosphate buffer and 1 M Na2SO4 in phosphate buffer
gradient mode: linear, starts with 1 M salt ending with 0.2 M salt in buffer
gradient time: varies 10-15 min, I dont see any improvement with differnt time.
The protein is lysozyme.
HPLC is from Waters 1525 Binary Pump.


Can anyone tell me what is wrong? Why do I get peak that is broader than isocratic peak?

Thank you so much

[danko]

I normally work with isocratic HIC of proteins

HIC is not an isocratic technique!
Actually you're describing the conditions as a gradient elution.

So, let's have more consistent description and more detailes - please.

Best Regards

[Uwe Neue]

What was the retention factor of the lysozyme under isocratic HIC?

What is it now under gradient HIC?

What was the sample solution used for injection (salt concentration)?

What was the amount of lysozyme injected?

[qrys]

Danko, Im sorry for messy posting, I corrected my original posting. Im just beginning to work with HPLC, please let me know what additional info you need. Flow rate is 1 ml/min.

Uwe,

tR for lysozyme under isocratic vs [Na2SO4] in phosphate buffer
0.67 min [0.2 M]
0.84 min [0.5 M]
1.3 min [0.7 M]
2.1 min [0.8 M]

I used t0 as 0.67 min as at 0.2 M salt is the condition when lysozyme is the least retained.

that gives k = 0.25, 0.94, 2.1 respectively for the above data

For gradient result, linear gradient 1 M to 0.2 M for 15 min, the tR is 8 min, im not sure how to calculate k for gradient.

Sample is 5.5 mg/ml in pure water, 5 ul of injection volume.

thank you

[Uwe Neue]

Was the flow rate the same for the gradient runs and the isocratic runs? Then the retention times are enough for making a guess at the gradient k.

Did you run the sample with the 1M salt? Too retained?

I can't find your flow rate. Assuming that it is 1 mL/min, then your lysozyme may or may not be unretained in 0.2 M salt, but at least it is close to unretained at the low salt concentration.

It could be that the peak is wider than it should be because you are injecting it in water, and not in high salt. The sample will migrate through the column a bit and smear out, until the high salt concentration has caught up with it. Try to inject it in a higher salt solution (I know, the solubility goes down, but that is a fact of life with this type of chromatography.)

[qrys]

Uwe,

the flow rate for gradient and isocratic runs is 1 ml/min

I ran the 1 M salt, too retained and I obtained a very flat peak.

I'll try using higher salt concentration for the sample.

Thank you

[danko]

Hi Grys,

As I mentioned earlier, HIC is a gradient elution technique and since you’re new to chromatography, you might like to work with more traditional setups – in the beginning.
So, for a start you’ll need to go for K>1 and preferably between 2 and 8.
The way you might approach the task is as follows:
1. Dissolve your sample in a salt containing solvent that is no more concentrated than the initial mobile phase concentration, but just enough diluted to keep the analyte perfectly dissolved.
2. Set the initial mobile phase salt concentration to a level that will capture the analyte at the head of the column. 1 M sodium or ammonium sulfate is usually adequate, but if not then increase the salt concentration to say 1.2 M.
3. Then immediately after loading the sample, start the gradient, diluting the initial concentration to no less than 0.5 M (in your case) with 0.5 – 1 % per min. Then increase the dilution (gradient) more rapidly until you reach 0 M sulfate in 3 – 4 min in order to wash off potentially strongly retained compounds.
4. If the peak/s is/are still too broad then add some organic solvent (5 – 10 % iso-propanol for instance) in eluent B and potentially in eluent A as well.
5. Once you’ve achieved a reasonable retention of the main peak (f. ex. 10 – 15 min) you’ll have the necessary experience to tweak the elution so that you’ll obtain a relatively professional chromatogram – meaning good peak shape and even more importantly a separation of potential impurities or/and degradation products.

A general precondition is to keep a constant pH in the region of the neutral pH (6 – for a start. That could be done by using some diluted buffer (ca. 0.05 M phosphate for instance) in both eluents i.e. A and B.

As for the background for the above suggestions (if interesting and it should be) you might like to read some literature dealing with the theory of HIC. It’s a longer story than I’m in the mood for writing about in detail.

Good luck and please get back and tell about the progress and potentially ask for additional suggestions.

Best Regards

[HW Mueller]

What charge does lysozyme have near neutral pH? If it is negative you will have to consider charge exclusion at low ionic strength and inadvertent pH instability.

[danko]

Hi Hans,

Proteins’ net charge is usually near zero in the neutral pH region. Besides, the relatively high salt concentration will take care of that problem.
From the info given, I would expect the main peak to elute nicely at 0.6 – 0.5 M salt concentration, so I don’t expect too many problems in this context.

Best Regards

[Uwe Neue]

The isoelectric pH (called the pI) of lysozyme is 11.0. At this pH, the predominant ionic form of lysozyme has a net charge of zero...

http://www.usm.maine.edu/chy/manuals/Bi ... Lys02.html

[HW Mueller]

Ok, so he only has to control ion exchange and adsorption.

Dancho, seems that I saw that he also works with much lower ion conc., and also injects in H2O. With radioactive simple ions (Na+, Cl+, . . . .) I got extreme effects due to this sort of mismatch on Atlantis Silica HILIC and ZIC-HILIC.

[danko]

This particular protein’s pI is 11 (thanks Uwe, I didn’t check it) – obviously one of several exemptions (the deviation that validates the rule). But the majority of all known proteins’ pI is in the neutral region (pH 6 – 8 ).
So, it’s predominantly positively charged at pH 7 for instance. And if the salt concentration in the eluent is low (much, much lower than I suggested f. ex. 0.1 M) some static interaction between the positively charged amino- acid residues and negatively charged silanol groups might occur. But the main part of interactions in HIC is contributed to the hydrophobicity that is induced by the very same salt that eliminates the ionic interactions. So here one might successfully make use of the saying: To kill to birds with one stone J
Hans, I - as well as Uwe - thought of the sample solvent (water) as a possible error factor, hence my suggestion to dissolve the sample in a salt containing water (1. suggestion). But honestly; I would be surprised if 5µL injection volume caused too much trouble.
You mention some considerable effects of pure water when used as a sample solvent in HILIC and I would expect that, especially when the injection volume is somewhat larger than 5 µL, but I’m not in the vicinity of believing that the current situation is due to the solvent type – even though I suggested one possible precaution i.e. salt in the solvent/water.
I think we are dealing with salt deficit due to the rapid gradient from 1 M down to 0.2 M salt.
The reason for not seeing problems (even though the very short retention times are troublesome enough) under the isocratic attempt is simply no or very poor retention, which in turn means no band spreading. Unfortunately, these almost unretained peaks equal no separation i.e. no need to use a column in the first place

Best Regards

[qrys]

Thanks to Danko, Uwe, and HW for your insights on this issue.

Danko, I tried your suggestion, I ran a gradient from 1 M to 0. 5M Na2SO4 for 1% per min (which I assume it takes 50 min) for the gradient then I ran steeper gradient from 0.5 M to 0 M for 5 min to get rid of the potentially strongly retained protein. My buffer is phosphate buffer pH 7.

I obtained very wide peak, tR=20 min, peak width as I estimated from the baseline is 30 min. I tried to dissolve the sample in 1 M salt, but I couldnt get it to dissolve, then I lowered the [salt] to 0.8 M. I.e. I ran lysozyme in 0.8 M salt through column of gradient 1 M to 0.5 M.

One thing that I dont understand here is that by starting with 1 M salt and lowering salt concentration, I should be compressing the peak, not widening it.

I have not tried adding the isopropanol because I want to make relations between the gradient result and the isocratic result that I already have. Secondly, I still think it should be doable without the organic in the mobile phase. I wonder if I miss something.

[Uwe Neue]

Well, it's not reversed-phase chromatography. Look in my book in Chapter 13 on how it works. I have not yet done the homework to try to calculate the retention factor at the point of elution for HIC, which determines the peak width.

Only in RP, and in some exterme cases of ion exchange will you get a peak width in gradients that is independent of retention. The theory for RP and ion exchange has been worked out well.

If you are mathematically inclined, you can try to calculate the retention factor at the point of elution for HIC yourself, but if I were you, I would first look into the literature, if somebody has done this already.

[danko]

Maybe the column (the stationary phase) is not up to the task?
I wouldn’t consider PEG for stationary phase if I were to develop a HIC method.
Who makes these columns anyway?

Best Regards