How to get rid of severe matrix effect?

[Ruth]

Dear,

I developped a method for 2 basic compounds using LC-ESI-MS/MS. Separation is done on a C18 column using a high pH mobile phase (MeOH/ H2O 25 mM ammoniumbicarbonate pH 9.5).

I analysed urine samples (which are postmortems, so quite dirty) with mixed mode SPE (Bond Elut Plexa PCX). Recovery is good (70-90%), matrix effect is fine for one component (85%, where 100% is no matrix effect), but not good for the first-eluting compound (16-40%).

I tried following things, but they didn't improve the results.
* low pH mobile phase
* different gradients with the high pH mobile phase in order to separate the analyte from interferences, this didn't really help, exept when using a 30 (!) minute run...
* I lowered the flow rate from 0.5 ml/min to 0.25 ml/min.

I was thinking to change the column chemistry (maybe try a phenyl column) or changing column length in order to get more resolution...

Since I still have to purchase these columns, I was wondering if you could give me some advice about what could be the best option... Or are there other ways to avoid the matrix effect?

Thanks

[Uwe Neue]

Column length won't do you any good. Changing the column chemistry is also not terribly promising, since you did some homework with the gradient already. As a quick fix, I would try methanol versus acetonitrile. You will know immediately, if there is any promise. In the absense of a quick positive result, I would revisit the SPE and see if this can be improved.

[grzesiek]

the problem is due to poor resolution from other peaks??

[Ruth]

@Uwe Neue: Thanks for the suggestion, I will try acetonitrile instead of methanol.

@grzesiek: Peaks are nicely resolved, the problem is caused by some co-eluting interferences from urine/SPE-procedure... I was thinking to change the column in order to be able to separate the analyte from these interferences, which didn't work with the C18 column I am currently using.

[sassman]

What are the specifics of your SPE cleanup?

[Ruth]

I use following SPE procedure:

conditioning with methanol, water, phosphate buffer pH 6
sample application
wash with water, acetic acid, methanol
elution with CH2Cl2/isopropanol/NH4OH (78:20:2)

[Uwe Neue]

The SPE method could be improved a lot. Since you are dealing with basic compounds, I would use a mixed-mode packing (Oasis MCX). It gives you the opportunity to use two retention mechanisms and this gives you much cleaner results compared to what you have at this moment. As a first attempt, you would load the acidified sample, wash with acidified water, then with methanol, and then elute with methanol with ammonia. The method is an excellent starting point, but it can be refined further, if the results are not satisfactory.

[Ruth]

The Bond Elut Plexa PCX SPE cartridges I am using are also mixed-mode SPE.

I use the CH2Cl2 as elution solvent since it evaporates more rapid than methanol. Recoveries are satifying, but unfortunately also a lot of interferences seem to be in the extract. Do you think changing the elution solvent can avoid this?

[Uwe Neue]

My team and I have developed all of the early extraction procedures for the Oasis product line. So I do have a bias for Oasis packings, and I know little about the Bond Elut Plexa material.
It could very well be that part of your interferences come from the use of methylenechloride. Methylenechloride will wash out large amounts of impurities, i.e residual polymer, from the SPE cartridge. This residual polymer can cause strong MS interferences. You would see it in a blank extraction with a UV detector, and you can test yourself if this is the problem by carrying out a blank extraction procedure (without any analyte) and monitoring the chromatography of this extraction with UV detection. If you see a large interference peak around the elution time of your analyte, then we know the source of the problem, and we can find a solution.
With respect to the actual procedure, I need a few additional details. I do not understand your wash protocol (I am sure that you do not wash with pure acetic acid). I also do not understand why you load basic analytes at pH 6, instead of at a more acidic pH.
If you want, we can discuss this off line in more detail.

[Ruth]

A good idea to test the methylenechloride.

About the SPE-procedure: the wash is first water, then 1M acetic acid and methanol. I could indeed use a more acidic pH for loading, which buffer/acid do you suggest then?

I am very interested in your opinion about this! Thanks a lot!

[Uwe Neue]

My standard protocol for bases on Oasis MCX start with acidifying the sample with acid, commonly phoshoric acid. The idea is to get the analytes into a known and defined charge state. Since you apparently do not have a problem with the loading, your bases are probably rather strong bases, and this is the reason why it works.

Let me know how you make out with the UV test of the blank extraction. If you see a lot of interference in the right retention range, we take it from there. If not, the problem is somewhere else, and we need to think about something else.

[Ruth]

I know the pKa of one of the components, this is ~8.5, so that's why I used a phosphate buffer at pH 6.

I will perform the test with methylenechloride next week (this week I am not in the lab) and inform you about the results. On my LC-MS/MS system there is no UV-detector. So I was thinking of doing following test to check the interferences in the elution solvent: I would evaporate some elution solvent and then resolve with the standard solution with compounds. I would compare this with the standard solution. I think this will give the same information as the test with the UV-detector you suggested?

Thanks!

[Uwe Neue]

OK, this will work also. If the results are identical between the two solutions, then the problem is not the methylenechloride elution and interference from the sorbent residues, but true interferences from the urine, and we need to improve the wash protocol.

[Ruth]

I just finished some experiments:

* I changed the mobile phase: I used the high pH mobile phase with acetonitrile instead of methanol and I used a low pH mobile phase (water with ammoniumacetate, pH4; methanol). In every situation severe matrix effect for the first eluting component was still there. I don't think changing the mobile phase will make the difference...

* I evaporated some elution solvent and resolved with standard solution. No matrix affect was seen, so interferences from the elution solvent (CH2Cl2, isdopropanol and NH4OH) don't seem to cause the problem.

* I also performed a SPE on spiked water instead of urine. A very limited ion suppression was seen here ( 10%), so much much better than working with urine...

As a conclusion: it seems to me that the interferences are derived from the urine. I guess improving the wash steps can decrease the matrix effect...

Please let me hear your opinion and your advice about improving the wassteps.

Thanks!

[Ruth]

Another idea: maybe I can change the elutionsolvent in order that the interferences don't elute, but only the compounds...?