Page 1 of 2

retention shift in only in plasma matrix

Posted: Wed Apr 11, 2007 2:16 pm
by neuger
can anyone explain this? dozens of injections in neat sol'n can produce fairly consistent retention times with a gradient of 2% - 14% ACN (across 6 min) with 0.1%TFA being the other mobile phase, yet the moment simple precipitation w/ 4 parts MeOH to 1 part plasma, evaporate supernatant and reconstitute with MP and inject, the retention moves almost 2 minutes at times and doesn't even really stabilize. This is such a simple precip method, I just don't get it!! Can such simple solvents like MeOH change the chemical makeup when they are used for precipitation of proteins, never seen that before? Could evaporation and concentration be a problem, again never seen it be a problem? If so, that is bad for me, I need it to achieve the sensitivity for the clinical samples we are testing.

I am pretty sick of this happening...here are the particulars:

YMC-C18 hydrosphere @ 30 degrees.
A: 0.1% TFA in H20. B: ACN w/0.1% TFA (I have tried with and w/o the TFA in ACN, it doesn't matter!)
reconstitute w/ 90:10 (A/B), inj 50ul.

retention in neat sol'n = approx 9.1 min
retention after reconstitution of supernatant = shifts from 9 - 8.5 - 7.1, etc

a large wash out step has been added to 40% organic at times and the run has been as long as 40 min to attempt to acheive equilibration, but none of this has helped...and you all thought I wasn't washing and re-equilibrating after a plasma matrix...lol, but seriously, anyone think they know what's up here??

Posted: Wed Apr 11, 2007 3:55 pm
by juddc
Bunch of questions:

What happens if you inject a standard immediately following one of your samples? Does the RT revert to original or remain low? Is the standard also in 10% ACN? What happens after you've injected a number of samples? Does the RT continue decreasing?

Something in your matrix could be coating your column. Have you tried flushing the column with a higher level of organic (80-100%) while monitoring your baseline?

Alternately with your MP starting at 2% ACN, your injection volume could be too high for a sample containing 10% ACN and the ACN in your sample could have an effect upon the MP composition. What are your system volume / column dimensions? Have you tried reducing your injection volume?

Posted: Wed Apr 11, 2007 4:09 pm
by neuger
it pretty much comes back to the 9 minute mark if we go back to pure standard. We have injected 3 more pure standard so far today, 9.1 each time. also, I forgot, we are only injecting in 0.1% TFA in water, regardless of pure standard or reconstitution of supernatant so the ACN situation can likely be ruled out, but I follow your logic. the column is 2 x150mm. also, I should tell you we are evaporating almost 2mls of methanolic supernatant and then reconst with only 60ul of the 0.1%TFA to inject 50ul. These are very small amounts of drug, ng amounts, that is why we are trying to concentrate so much so I don't see any overload. My heights are less than ten on the y axis with UV detection, but perhaps we are overloading with any remaining plasma components and do need to wash up to 80% organic??

Posted: Wed Apr 11, 2007 4:55 pm
by Mark Tracy
have you checked the pH of the sample after reconstituting it?

Posted: Wed Apr 11, 2007 5:56 pm
by neuger
how? Very difficult to do that, I don't have a probe that will make much contact with that small an amount and if I dilute with more, that is not an equal test of what I am using, any suggestions. I again understand this concept and this is a possibility, acids above their pKa shifting to earlier retention time, but 0.1%TFA is pH of 1.9. why would it move over time, if the pH was different, shouldn't the retention be reproducible in plasma and not slowly reach where it may want to be, ex: 8.5 in one inj, then 8.1, then 7.5, then 7.0, etc, etc from the plasma samples?

Posted: Wed Apr 11, 2007 6:19 pm
by juddc
Interesting...

If you're not using solvent in your reconsitution, then solvent effect is out, so it sounds a bit like junk in the trunk to me (absent a pH issue).

A thought: Could you clean up your MeOH extract prior to drying down? An SPE protocol might help, I think - something like the following:

Precipitate with MeOH as usual. Pull off supernatant and either dilute a bit with water or run neat down a properly conditioned SPE cartridge to hang the junk up (and pass your analyte). Dry the SPE eluate, then reconstitute and inject as usual.

Alternately, you might use SPE to concentrate your analyte and clean up your sample simultaneouslyif you find the reight sorbent and conditions, but that would take some homework to get right.

Are you using a guard column?

Posted: Wed Apr 11, 2007 6:30 pm
by neuger
yeah, funny you say that...we tried that before with something else and it worked for the most part...straight up SPE would be very, very tough, we've tried...highly hydrophilic compound, maybe there is just some junk in plasma causing it, that is what we have theorized the whole time, but we hoped someone thought it was something simpler related to chromatography, pH, etc.

Posted: Wed Apr 11, 2007 8:42 pm
by JJG
Have you tried conditioning the column with up to 10 injections of plasma extracted samples, and then seen if the retention time stablizes?

Another thought would be to take the neat standard and extract it as though it were a plasma sample. If it had the neat ret. time, you would know it is definitely from the plasma, not the process.

Posted: Wed Apr 11, 2007 9:25 pm
by Uwe Neue
You are carrying a fair amount of protein to the column with your method. MeOH precip is worse than MeCN precip. SPE would be a much better approach. My bet is that the problem is in this area, since you get consistent results when you just inject the standard.

Posted: Thu Apr 12, 2007 6:45 pm
by neuger
Image

some visualization that may help better anyone's advice. here is the result of 4 consecutive injections in plasma matrix, everything is consistent with what is described in the original post, I have included the gradient steps...FYI, look close, both peaks of interest are shifting earlier and earlier, not just the one around 9min that is obvious. also, we have experimented with different run times to achieve re-equilibration after the washing step in the gradient and it hasn't seem to help.

Posted: Thu Apr 12, 2007 7:36 pm
by juddc
What I see in your chromatogram reinforces my feeling that you've got junk on your column - you're losing efficiency and the peak at about 10 min is moving in as well, I see.

Quick experiment alluded to by Dr. Neue: Try precipitating with MeCN rather than MeOH.

I think this one can be solved with a well designed sample cleanup procedure. The fact that you're looking at a hydrophilic analyte and using a mobile phase with high water content exacerbates any problems you might have with junk on an RP column.

If you can't clean up the sample, you might consider alternate separation modes (an amino column comes to mind) which might be more forgiving. I know...lots of work, but it's a suggestion.

Posted: Thu Apr 12, 2007 8:03 pm
by neuger
Major solubility problems in acetonitrile...the recovery is terrible, but I think I follow the chromatographic situation you present.

Posted: Thu Apr 12, 2007 11:40 pm
by Bryan Evans
Hi neuger -

If you contact me , I might be able to help. We have a new product
that might be helpful to you - thank you.

Posted: Thu Apr 12, 2007 11:47 pm
by Uwe Neue
Here is a quick SPE procedure. I am sure that one can design a better one, but I would need to know more about your analyte.

Acidify your sample with HCl or similar.
Put it through an Oasis HLB cartridge.
Wash with a bit more HCl (how much depends on the size of the cartridge that you choose).
Elute with about 30% MeOH.

Suck dry, redissolve, inject.

It's some work, but proteins are washed through in the wash, lipophilic stuff like phospholipids stay behind, and your sample is much cleaner.

Posted: Fri Apr 13, 2007 3:27 am
by james little
Really strange. The other thing you might consider is the presence of the phospholipids. Those still hard to explain by their presence.

The acetonitrile doesn't get the phospholipids off the column very well.

However, I found that even injecting 50-100 samples with phospholipids on column did not affect peak shape or retention time significantly.

see http://users.chartertn.net/slittle/ section of lipids matrix effects in LCMS..