Polymyxine B sulphate variable retention time

[aceto_81]

Hi,

While working on Polymyxine B Sulphate, according to the Eur Ph method, we discovered that the retention time of the peaks is really unstable. During the sampleset, we can see the retention time is sometimes increasing for a while, and then suddenly decreasing.
First we were thinking that it would be an temperature problem, but after putting the column in a oven set at 30°C, the same issue appeared.
Next thing to know: there are other compounds also injected, and the variable retention time only occurs with Polymyxine, not with the corticoids.
We already tried to switch the method to UPLC (to shorten the run, but also to overcome the problem), but the same problem occurs here.

Method:
4.46g Sodium Sulphate anhydrous / l, brought to pH 2.3 with phosphoric acid. This is mixed with ACN in a composition of 20% ACN and 80% sodium sulphate solution.
Flow 1.0 ml/min (0.4ml/min for UPLC)
Detection 215 nm
Column: Supelcosil ABZ+ 250 x 4.6 5µm (BEH Shield RP18 100 2.1 1.7µm for UPLC)
Column temperature 30°C.

Already tried to put a buffer in the mobile phase, instead of the sodium sulphate, but then the polymyxine looses retention, but also has a variable retention time.

Any ideas?

Ace

[Uwe Neue]

Are you getting the same changes on both recommended columns?

[aceto_81]

Uwe,

The recommended columns are Supelcosil ABZ+ , Uptisphere ODB and Symmetry.

We tried the Supelcosil ABZ+, and afterward on UPLC the BEH Shield RP18 with the variable retention times as a result.
As we already tried these 2 columns with the same result, we actually didn't suspect the column.

Ace

[Bryan Evans]

Below is a method for Polymyxin B on Cadenza CD-C18 (12nm pore size) and Cadenza CW-C18 (30nm pore size):

http://www.imtaktusa.com/site_media/fil ... TI427E.pdf

Just curious, does anyone know why the Eur Ph method was developed with sodium sulfate instead of NaH2PO4?

[Uwe Neue]

The polymyxin has multiple amine functions, which might make its retention extra-ordinarily sensitive to the development of silnaols as the column ages. I suggest to use the XBridge or UPLC BEH C18 column instead of the one with the embedded polar group that you are currently using.

[HW Mueller]

Is the BEH C18 column mentioned by aceto and Uwe the same material?

Since this compound, if as described by Uwe, should be very pH sensitive it might be helpful to use a good, strong buffer, but keep the Na2SO4 for better ion strength control.

Just a side remark: It would be nice if the original poster gave descriptions of their analytes so that some (many?) of us wouldn´t be forced to look them up before being able to form an opinion.

[Uwe Neue]

Hans: no, the two materials are different. The BEH Shield RP18 is a monofunctional bonded phase with an embedded polar group. The BEH C18 is a trifunctional C18. Both packings are based on the bridged-ethyl hybrid (=BEH) base material that gives the packings good stability in the alkaline pH range. What counts at acidic pH is - among other things - whether the packing is bonded based on a monofunctional ligand or a trifunctional ligand.

The ABZ packings have their own complications.

[grzesiek]

ok so if you change the column to BEH C18 then you are no longer using Ph Eur method so why not try USP method?

Column is L1 (the same dimensions), detection 212 nm, flow the same

Mobile: 0.1M tribasic sodium phosphate : ACN 77:23, adjusted to pH 3.0 with phosphoric acid

So, method looks simmilar, only phosphate buffer is used

[Uwe Neue]

If my interpretation is correct and some of the problem is associated with the multiple positive charges on the analyte, I do not recommend the USP method: the ionic strength is way to low to make for an acceptable method. I do not know how the EP methods work. I believe they only suggest the packings that are supposed to work, but give you freedom to use a substitute. My substitute, the XBridge C18 packing or its UPLC equivalent, is most likely the best choice considering the problem that you have.

[aceto_81]

Is the BEH C18 column mentioned by aceto and Uwe the same material?

Since this compound, if as described by Uwe, should be very pH sensitive it might be helpful to use a good, strong buffer, but keep the Na2SO4 for better ion strength control.

Just a side remark: It would be nice if the original poster gave descriptions of their analytes so that some (many?) of us wouldn´t be forced to look them up before being able to form an opinion.

Thanks for al the suggestions.

@Hans: I tried with a 50mmol phosphate buffer at pH 2.3, which gives a lot less retention. Afterwards I added the sodium sulphate to the buffer and did my runs again: the retention was back, but so was my problem....
About your suggestion for the description of the analyte:

See also: http://en.wikipedia.org/wiki/Polymyxin_b_sulfate

Also tried 0.1% TFA in a gradient from 5% ACN to 95%, but the shifting also occurs.

Uwe, I will try the BEH C18 with the original mobile phase today. Keep you informed.

Ace

[aceto_81]

Uwe, I tried the BEH C18 with the original mobile phase (sodium sulphate @ pH 2.3), with a flow of 0.2ml/min, but still I have the changing retention times (about 0.4min decrease per 40 minutes)

Ace

[HW Mueller]

Did the 50mmol buffer also give shifts? Did you have corticoids in all experiments were you saw shifts only with the polymixine? What are the retention times of the corticoid and the polymixine?

[aceto_81]

The retention time of the polymyxine B on the UPLC is about 8.6 minutes for the main peak. The corticoid is about 6minutes.
(Method with 50mM phosphate buffer + 4.46g/L sodium phosphate, BEH Shield RP18 flow 0.4ml/min)


Ace

[Uwe Neue]

I am puzzled. It does not seem to be a deterioration of the packing, since all the packings that you tried have the same problem. The mobile phase is fine, the buffer concentration and the pH are adequate. You have temperature control…
I have seen some analytes to be more temperature sensitive than others. Do you have the 3.5 micron version of the XBridge BEH C18?

[aceto_81]

We don't have the X-bridge, only the UPLC equivalent.