variations in repeat injections of high % organic samples

[MHANLEY]

I'm running a microsome stability assay in which the samples are quenched in MeOH giving a [final] of 66% v/v. The samples are then centrifuged to remove ppt then loaded onto a polypropylene 96 well plate and sealed with 'Zone-free sealing films' from Alpha. They seal the plate well and I'm happy I'm not getting evaporation but I've noticed quite an annoying side effect. I perform two runs/injections per sample and I've noticed that on average the first injection gives much greater peak areas than that of the repeat injection (for both test article AND internal reference), sometimes twice the value. The ratio of the peaks however remains consistent.

If the sample contains no volatiles I do not get this issue, with duplicate injections giving very similar values.

I have performed 5 injections from the same well with samples containing the same MeOH content as my assay samples and I do not see a steady decrease in peak area with each subsequent injection, just a difference between the first injection and any of the subsequent injections

I only load 7 ul and the repeat injection is always performed within a maximum of 30 min (minimising any issues associated with sample evaporation) so I'm happy that the issue is not due to depletion of the sample between injections.

Ideally I would like to pierce the film, withdraw the needle, wait a while THEN take the first injection. Unfortunately the closest I think I can come to this is to perform a zero volume injection but considering the excruciatingly slow pace of the needle carriage (this is an old Alliance HT 2795!!) this would mean increasing my run time from around 20 hrs to 30 hrs per assay, significantly affecting my throughput.

As I'm looking at changes of [test article] as measured by ratio to the internal reference I can actually extract the results I want with a high level of consistency and accuracy, I just hate seeing such wide error bars.

Sooooooo..........Has anyone else observed this before, have any ideas what might be causing this variation, and perhaps have a solution which wouldn't increase my run times by up to 1/3.

Thank you

[Peter Apps]

Can you eliminate the possibility that some of the analyte and standard are coming out of solution ? Are you by any chance using a cooler on the autosampler ?

Peter

[MHANLEY]

Hi Peter,

Compounds are at 2 uM (maximum) in the assay samples which contain 0.1% DMSO and 0.3% MeCN to aid solubility without perturbing the assay. The samples are quenched into ice cold MeOH ([final] = 66% v/v) with the samples being held at 20 deg C within autosampler.
The Log Ds are all between 1.5 -3.5 so at this concentration I expect them to be soluble and along with the high [MeOH] I'm fairly sure the compund isn't dropping out during the sample prep or whilst in the autosampler. Also this behaviour is consistent across all compounds.

...........but, it would be folly to say with absolute certainty that this is not the case. Are you suggesting that piercing the film and releasing vapour pressure may have an effect on the solubility of the compound?? This then would (seemingly) always have to have a equally proportional effect on the precipitation of both the test article and internal reference. What are you thinking??

[Peter Apps]

When the likely culprits are not guilty, then something less likely must be happening:

With a high proportion of methanol, evaporation of the sample is possible - but would give larger peaks (not smaller) in subsequent injections, and the peaks would keep getting bigger through a re-injection series.

Carryover would make the peaks bigger.

Oxidation due to exposure to air could make the peaks smaller, but they would get smaller and smaller throughout a sequence.

Precipitation could well affect analyte and IS the same if you are choosing an IS which is chemically similar to your analyte.

Is it possible for anything to partition into the sealing film ? What if the needle punches a little bit of film out and drops it into the well ?

What happens if you manually stab a hole through the film before you load the plate into the autosampler ? Do you then get smaller peaks with the first injection ?

The simplest explanation for all the peaks getting smaller in equal ratio is dilution of the sample (presuming that the injection volume is stable). Is it possible to measure fluid volume in a well immediately after sampling (suck it out and weight it would be my suggestion) ? Dilution could only happen if the autosampler was adding fluid to the well.

What is the run schedule - do you do the two injections from one well sequentially, or do you do one sample from every well and then go back to the first well for its second injection ?

Peter

[MHANLEY]

I've checked the hole punched by the needle and the plastic is merely displaced rather than detached.

When troubleshoot this issue I've watched the syringe pump and cannot see any movement other than withdrawing the 7 ul into the needle, but just to make sure I could at some point check the volume post 1st injection.

I could try pre-punching a hole but in different positions over a number of wells to see if the anomaly is a function of the needle actually punching through the seal?

I cycle every four samples and my run times are roughly 4.5 min so no more than 18 min between 1st and 2nd injection from any one well.

Thanks for the ideas.......I have a terrible feeling I might be scratching my head for a while to come!!!!

[Peter Apps]

I cycle every four samples and my run times are roughly 4.5 min so no more than 18 min between 1st and 2nd injection from any one well.

OK, this establishes that it is something happening in the well, only after an injection, if it was something to do with autosampler operation, or it was happening in the wells before injection, then you would get a low result for wells 2 - 4. Try doing the two injections form each well consecutively and see how it looks.

Peter

[James_Ball]

If you are losing the methanol, it would evaporate at a slower rate the closer the ratio of water to methanol becomes. You quench with methanol, would the reaction begin again if the methanol to water ratio becomes lower?

Assuming stainless steel needle, could the metal act as a catalyst?

Just throwing out some wild ideas since most of the obvious ones don't seem to be in play.

[carlo.annaratone]

Could it be that some sample is condensing on the film, so it's carried over by the needle in the first injection, but not the subsequent ones?

[Peter Apps]

Since you are transferring prepared sample to the 96 well plate (rather than doing reactions on the plate) maybe it would be worth trying ordinary sample vials with septa caps instead.

Peter

[MHANLEY]

Thank you to everyone for your suggestions. Haven't had chance to investigate further but I'll definitely look into some of these ideas.

Carlo....I've not thought too much about carryover as the hole created by the needle in the film is no larger than the needle itself and I presume wipes any large drops off as the needle is withdrawn. I haven't taken out and had a look at the injection port so I'm not familiar with how tight the needle fits and how significant any carryover might be....enough for up to 2 x variation?? However seeing as I don't see this effect with aqueous samples (the high surface tension reducing the amount of sample left on the needle after withdrawal through the tightly fitting hole on the film) this migh well be the case?!.....wait a second, sorry a small eureka moment............even a small volume of MeOH on the outside of the needle (with a lower surface tension) might flow to the end of the needle and be pushed into the injection port ahead of the needle. As there is a 'purge volume' also injected into the loop and not just the 7ul of the injection, any volume of carryover would be injected. Ok.......perhaps you could be onto something here. Only question is how to check this out...I think the needle wash occurs AFTER every injection. I might try and see if I can wash before hand....without significantly diluting the sample at the relatively wide bore of the needle end!!

Peter....unfortunately cost and time prevents me from transferring samples to individual vials (Assay is perfomed in 96 well format)

James....The MeOH quench precipitates the enzymes required for metabolism. I've looked at evaporation rates after puncturing the well and within the period of duplicate injections the change in volume is negligable.

Again thank all you for time!!