LC MS MS: splitted peak after extraction!

[Fluterd]

Hello,

I request some suggestions for a problem in my LC MS MS method development for a basic drug and its main metabolite.

I've optimised a mobile phase, column, LC and MS (MRM) parameters which all work fine when I inject 5 µl of the standard dissolved in starting conditions of the mobile phase. So, it seems to me that my parameters are fine and the system is working well.

I extracted some urine with added standard with SPE and reconstituted the samples in mobile phase. When I inject angain 5 µl, the peak of the metabolite is split. This is seen for the two MRM transitions that I am using for that compound... I have tried some different SPE columns and different SPE protocols, however, all of the used methods show some level of peak splitting.

Has somebody experienced the same problem? It's hard to find any literature about it...

Thanks in advance for your help!

Ruth

[Uwe Neue]

There are multiple possibilities, but we need more information to exclude possible phenomena. Please describe the mobile phase, the gradient, the final extraction solvent from the SPE, the nature of the analyte that is giving the problem etc. All of these items could be causing a problem of the type that you are describing.

[Fluterd]

Hello,

The mobile phase is (A) water 25mM ammonium formate (B) acetonitrile with 0.1% formic acid. The run is isocratic at 80% B.

The compound is a polar metabolite, containing an amine group.

I tried different SPE elution solvents: methanol, ethylacetate or isopropanol as organic compound in combination with or without ammonium hydroxide (these protocols were found in literature and worked fine there for these compounds...)

Thanks!

[Uwe Neue]

Do I understand you correctly that you do not have any formic acid in the aqeous portion of your mobile phase? If that is the case, I would change that first and add formic acid to the aquoeus mobile phase.

Next I would try to dissolve the sample in a solvent with less organic, about 50%/50% aqueous/organic. You need to have the buffer in this as well.

If this does not help, we need to have a closer look at the details of the SPE procedures, and/or at the details of your sample injection. Since you are running isocratically, I do not foresee a problem with the injector.

The other thing that bothers me about your LC procedure is that you use such a high concentration of organic. This is unusual. Either your analyte is very hydrophobic, or it is essentially unretained. What is the retention factor? Or, what is the retention time at what flow rate, and what are the column dimensions?

[Fluterd]

Hello,

My mistake: I forgot to mention that the aqeous mobile phase is set at pH 3 by adding formic acid.

I use the exact mobile phase as injection solvent because I've tested several conditions (including the 50/50 you mention) and 80% appeared to be best. This was however tested with the standard, not with an extracted standard. Do you think this could make a difference?

I know this is a highly organic mobile phase. I've planned to check for matrix effects and then -if necessary- apply a longer gradient. The highly organic content is just a very short method that I am using to test different conditions. The analytes are eluting between 1min and 1.5min. The column I use is 3.0*100mm, 2.5µm particels and the flow rate is 0.5 ml/min.

Thanks for your help!

[Fluterd]

Sorry, small mistake about the column diameters. The column is [b][b]2.1*100 [/b]mm[/b], [b]1.9 µm[/b] particles. Flow rate 0.5ml/min. Elution of compounds between 1min and 1.5min.

[Uwe Neue]

In the column mentioned first, your analytes would be essentially unretained. In the second case, you have a bit of retention, k' between 1 and 2. The latter is usually OK.

Your injection volume is OK. I think that you should try nevertheless to inject the extracted standard in the 50/50 mixture, with some formic acid. It is possible that the ammonia in your extraction solvent leads to a pH change that together with the low retention factor creates conditions for the double peak. To free the analyte from the potential residues of the extraction solvent, you will be better off either working at a larger retention factor or using a more polar injection solvent to concentrate the analyte at the column top.

[Fluterd]

Hi,

I've tried some of your suggestions.

Changing the injection solvent didn't solve the problem.

I've changed the isocratic elution conditions to a gradient, and this solves the problem. Some irregular small peaks are seen at the beginning of the run (I suppose this are the same interferences that caused the splitted peaks, but they are now separated from the metabolite), and the peak of the metabolite is eluting a little bit later and has a nice shape.

Can I correctly conclude that the extraction procedure produces some interferences with the same MRM transitions as the metabolite and that they can be separated from the metabolite by applying a gradient?

Thanks for your help!

[Uwe Neue]

I am not convinced. I would take the same MRM transitions as proof that both peaks are the same compound. As matter of fact, I have seen two different retention times for the same compound when it was injected from a strong solvent into a gradient: one at the standard retention, and one close to unretained. You need to get to a point where you inject the sample in a way that does not cause two peaks for one analyte.

[Fluterd]

I've used the starting point of the gradient as injection solvent (i.e. 20% aqeous, 80% organic), so I don't believe teh injection solvent is too strong.

Maybe I didn't explain well enough the phenomenon I have seen using a gradient.
When using the gradient, I have a nice peak of the metabolite at a certain retention time (cfr. when I inject standard), but before this elution there are some very small peaks. However, the two MRM transitions do not peak at the same moment then. This is why I supposed these very small peaks are caused by the extraction and are not the metabolite...? If it was the metabolite I would see the two MRM transitions peak at the same time, no?

Thanks, I really appreciate this discussion!

[Uwe Neue]

How about the pH of your sample? This can also cause strange elution patterns.

Anyway, when we started, you had a huge second peak with the same MRM. Now you have only small peaks with the same MRM. Is this correct?

[Fluterd]

Indeed, the first problem was the second huge peak for the two MRM transitions of my compound. This peak has now clearly dissappeared, but here and there, there are still some very small peaks, but not at the same time for the two transitions.
So, I concluded the problem is solved, but I would like to have a proper explanation for this phenomenon...

Do you mean the pH of the injection solvent or of the sample before extraction?

[Uwe Neue]

If the sample is injected in a solvent composition that either is too strong (e.g. in DMSO) or has the wrong pH (and you are using a weak buffer), then you can get two peaks from the same compound. Some of the analyte is eluting in the mobile phase with proper retention, some of it is eluting much earlier. In the case of DMSO, this has been explained by the high hiscosity of its mixture with water. You end up with a front and a rare barrier of high viscosity that prevents mixing of your analyte with the rest of the mobile phase. In the case of a strong departure of the analyte buffer pH from the mobile phase pH, you are running first a chromatogram at the pH of the injection solvent, followed by a chromatogram at the pH of the mobile phase. Usually, this second phenomenon does not result in two peaks, but double peaks and fronting shoulders, but if the details of the injection are not understood, it is tough to judge how the peaks will look like.

[Camisotro]

As matter of fact, I have seen two different retention times for the same compound when it was injected from a strong solvent into a gradient: one at the standard retention, and one close to unretained. You need to get to a point where you inject the sample in a way that does not cause two peaks for one analyte.

Just to add to Uwe Neue's analysis above... I work in a university MS research lab and I've been given some leeway to learn some chromatography basics through trial-and-error. The other day I was testing out various chromatographic conditions to resolve an analyte from two of its metabolites.

I observed that if I happened to pick a mobile phase that just barely retained my most polar analyte, sometimes I would see it as a split peak - some of it unretained, some of it retained.

Perhaps your optimized mobile phase was actually very close to that boundary condition, and your sample prep added or removed a matrix component that tipped it across the boundary?

Just a thought. Could be off the mark here and I'd appreciate any corrections

[Fluterd]

Thanks for your remark, but in my case, the split peak eluted later than t0, so there was retention of both parts of the splitted peak...