Growing pains with SPME, GC-MS and standards

[Peter Apps]

Hi Lue

As a biologist re-formatted into an analytical chemist I feel your pain.

An Rx 5MS column would probably not be my first choice for this, but from what I can see of the chromatogram you are getting adequate separations in terms of retention time. The tailing on the peaks could be due to a deteriorated column, or to active sites in the inlet. If the latter a different column will not help, but it would be covered by routine inlet maintenance that your chemistry collaborator should be doing.

The advantage of using a liquid medium is that the liquid phase could be homogenized by stirring or shaking when you do the analysis. An homogenous liquid phase will give more repeatable partitioning of the analytes into the vial headspace, which will give more repeatable absorption into the SPME fiber, and more repeatable peak areas.

Are you taking repeat samples from cultures ? - if you need to be able to sample again from the same bugs, then you cannot use "destructive" techniques like heating the sample to get more volatiles into the headpsace, or adding large volumes of a dilute internal standard and homogenizing the sample into that.

If you are growing the bugs in sealed vials you might run into the issue of accumulating volatiles starting to inhibit growth, or particular metabolic pathways. Also, are these things anaerobic ?, do you have a controlled atmosphere in the vials ?

Back to the issue of internal standards and repeatability. Using an internal standard is very useful if you have steps in the analysis where volumes are variable - variability in dilution or concentration steps are perfectly compensated for by an IS, and the IS does not have to be a very close chemical match to the analyte. If the analysis includes partitions between two phases then ISs can still be useful, but then they must be as close a match as possible to the analytes - which is why people use isotope labelled internal standards. SPME is a partition-based preparation technique, which is why your ISs would have to match the analytes, and with the diversity of analytes that you have getting ISs to match all of them is going to be difficult.

You need to know what the source of the variability is, and whether it is the kind that can be fixed by internal standards. From the look of the chromatogram the tailing is probably compromising the integration, and adding an internal standard will do nothing to help this. In fact it will make the repeatability worse because the poor integration of the analyte and the IS peaks will add to one another. So you need to present your chemical collaborator with replicate samples that you know are the same - to produce these you need your analytes dissolved in agar at concentration similar to what you get with the bugs growing on it. Make up 5 vials with this spiked agar in it and no bugs. Take a SPME sample from each and get them analysed and see how repeatable the areas are. If they are poor, make up a solution of the analytes in a solvent that can be injected to the GC and run five replicates of that. If the repeatability is still poor the problem lies in the instrument, not with the SPME sampling, and adding an internal standard to your samples is probably not going to do any good at all.

You probably already know that there is a lot of literature in the forensics area about volatiles from carcasses and cadavers - I can send you some papers if you ping me a peter-j-apps-at-g-mail-com. Take out the dashes and make at an ampersand , it keeps the bots at bay.

Peter

[spiderylu]

Hi again Peter,

I've had my head buried in exams and coursework for the past week or so, so sorry I haven't been in touch.

Using a liquid media could be useful - we're tracking the volatiles emitted during the active growth phase of the bacteria to get a sense of what they give off as they're metabolising. I suppose, based on your explanation, that a liquid medium might be useful for this.

We are growing the bacteria in capped autosampler tubes...we're inoculating with 5ul of 10^7 CFU/ml and sampling over a 24 hour period, so I don't think that accumulation of volatiles is much of an issue here, although it's an interesting thought....something I hadn't thought about. Proteus mirabilis is facultatively anaerobic, which I guess is why they're in sealed containers.

I will look at some of the papers on carcass volatiles and see if any standards are used for that. Other (cheaper) methods of using Tenax tubes or Carbopac have been suggested. I don't know much about them, what would you say about that?

Thanks!


Lue

[rb6banjo]

To use traps like you mentioned, you would have to have a flow of gas into and out of the container where your sample resides. For your system, the gas could not interfere with the growth of your bugs. The idea with the traps is that as the volatiles are generated, they are swept out of the vial by the flowing gas and through the sorbents (you mentioned some good ones above) where they are trapped. Then, the traps would have to be taken to a chromatograph equipped with the right plumbing to desorb the trapped analytes from the solid sorbent into the instrument for analysis. What you're looking for is a gas chromatograph set up for "thermal desorption" analysis. TD systems are quite a bit more complex than a SPME sampling system but they do work well. Precision can be much more of an issue with TD systems compared to SPME systems (with our without an internal standard).

[spiderylu]

Yes, one of the things I was worried about was a loss in accuracy using an airflow method. I suppose that might be up to my advisor...

[Peter Apps]

Hi Lue

I would go to adsorb - desorb trap sampling only if the issue with SPME was a detection limit that was too high. If you have poor repeatability with SPME you will probably not get any better results with adsorb - desorb traps.

I'm no microbiologist, but 5 ul innoculations sounds very small scale. Which is often good. What size of vial, and how much medium are you using ?

You really do have to check the repeatability of the analytical step by running standards before you start worrying about the sampling and microbiology.

Isn't Proteus and amoeba ?

Peter

[spiderylu]

It's 5ul bacteria on 4ml agar in a 20ml autosampler vials.

This method has been previously used and published by my lab - my advisor wanted to look at an internal standard because he'd been speaking to someone in industry. I think in the end he said there was only around 20% variability between samples, which I understand is not very much.

Proteus mirabilis is a gram negative, flagellate bacteria causes urinary tract infections, so I'm not really friends with it!

[Peter Apps]

It's 5ul bacteria on 4ml agar in a 20ml autosampler vials. I'm guessing that the agar gives a layer about 10 mm deep. The bugs grow only on the surface (or are you doing stabs ?). The more agar you have the more metabolites diffuse into it instead of into the headspace where you can get at them with SMPE, so if you use less agar you might get more repeatable results.

This method has been previously used and published by my lab can you send me a .pdf = then I don't need to keep asking dumb questions about microbiology !- my advisor wanted to look at an internal standard because he'd been speaking to someone in industry. I think in the end he said there was only around 20% variability between samples, which I understand is not very much The sampling and GC steps should be able to give about 2% rsd, if the 20% is just sampling and analysis than you need to have a "friendly chat" with your analytical collaborators.

Proteus mirabilis is a gram negative, flagellate bacteria causes urinary tract infections, so I'm not really friends with it! Yuk, considering where it lives I suppose the ability to swim upstream is a useful adaptation

Peter