mystery mass spectrum interpretation

[Peter Apps]

Hi all,

Can anyone throw any light on what compound this may be ?; GC-MS 70ev electron impact on a Scion single quad. If the small peak at 133 is the molecular ion it would fit with its retention index of about 2030 on a Restek wax column.

The peak is clean; the spectrum stays the same all the way across. It is from African wild dog urine, so possibly a metabolite, but with no hits in the NIST library.

https://drive.google.com/file/d/1qBOiiL ... drive_link

Thanks.

[JMB]

Peter,

Could you blow up the molecular ion region m/z 130-140
( possibly summing more scans, if necessary) to see the C12 / C13 ratio.

Same for the m/z 75 / 76 ion.

Regards,

JMB

[Peter Apps]

Peter,

Could you blow up the molecular ion region m/z 130-140
( possibly summing more scans, if necessary) to see the C12 / C13 ratio.

Same for the m/z 75 / 76 ion.

Regards,

JMB

Certainly: https://drive.google.com/file/d/1-gx-Sq ... drive_link

and

https://drive.google.com/file/d/1Xo2R0t ... drive_link

Thanks.

[JMB]

Peter,

This may be a small molecule, but it is nevertheless a VERY nasty one.

[A] IF the MW =133, then the only things that I can be definite about are,

1) N atom count ODD [1,3,5,...]

2) No Cl, Br, I

Looking at the magnified molecular ion region,

1) Relative Intensity (RI) m/z 134 to 133 is 124,000/2116000 = 5.8%, therefore C~5.

2) RI's decrease smoothly across m/z 133 to 137, EXCEPT m/z 135 which seems more intense than it should be----presence of S1 ??

[C] Inspection of the full-scan spectrum indicates possible presence of -OCH3, from m/z 133 ---> 102.

If [B; 1 & 2] true, then possible molecular formula is C5 H11 N1 O1 S1 =133 Da.
OBSERVED CALCULATED
m/z 133 100 100
m/z 134 5.8 6.7
m/z 135 4.7 4.7
m/z 136 2.1 0.2
You can see that the Obsvd/Calcd are similar, but not an absolute match, esp. m/z 136!!!
For C5 H11 N1 O1 S1, the Rings/Double Bond count = 5 - 11/2 +1/2 +1 = 1.

That is as far as I am at the moment.

Regards,

JMB

[Peter Apps]

Great stuff, thank you !

Minor deviations from calculated intensities could be due to interference from co-elutions; this is mammal urine and very complex. I am going to run some samples again on a 60m column rather than a 30m.

To an extent the "nastier" it is the better, as in an unusual structure, because what I am looking for is compounds that are specific to African wild dogs, and are therefore better candidates as a chemical signals than an "ordinary" compound that occurs in several species.

So far, and I haven't checked every sample, this 102, 103 compound occurs only in wild dogs and not in jackals, foxes, spotted hyaenas or any of the cats. So far it is the only wild dog compound that does not occur anywhere else. Which makes it very interesting quite apart from its probably "nasty" structure.

I sense that I will have to find a lab with a high-res MS. I can do CI, but nothing beyond integer resolution.

I do have a PFPD which would confirm the sulfur, and at lower sensitivity, the nitrogen.

Thanks again - any further thoughts would be welcome.

[JMB]

Peter,

Yes, HR will be a necessity for a new compound and will be very useful because at MW~133 the number of possible CHNOS combinations will be quite limited.

Running CI is a good idea, because it may help to “clean up” the molecular ion region, which you will need to have a definitive MW.

A good way to check the homogeneity of a chromatographic peak is to use mass chromatograms to compare the elution profiles of different ions.
If you were to plot, e.g.,
m/z 133; 134; 135
&
m/z 135; 136; 137on separate sheets, OVER THE SAME TIME INTERVAL, you MAY be able to visually tell if your chromatography peak is pure or a mixture of two components.
You have an interesting project!

Regards,
JMB

[Peter Apps]

Thanks again JMB.

I will look in more detail at the peak profiles for the molecular ions - there is certainly nothing in the large rpeaks that suggests a significant co-elution (which is very common with such complex samples).

My next analytical step is to run it on the PFPD.

I will post the results of that.

Peter

[Peter Apps]

The mystery deepens, or perhaps simplifies.

The results that I posted were from an analysis with a hot split injection. With a PTV injection the m/z 133 and 134 ions coelute with the main ions, but 135 and 136 elute in the leading edge of the peak; https://drive.google.com/file/d/1UqZ8u2 ... drive_link

and the spectrum for the molecular ion and isotopes at the peak of the main peak is then; https://drive.google.com/file/d/1JhcFwk ... drive_link

Some samples have the main peak but not the 135,136 peak; https://drive.google.com/file/d/1BUCYxO ... drive_link

and the spectrum is; https://drive.google.com/file/d/1sf3kPD ... drive_link .

Clearly 135 and 136 are not from the main component of the peak.

[Peter Apps]

Update; on a 60m column the m/z 135, 136 separate from 102, 103 and a library hit of 4-methoxybenzaldehyde looks plausible for the 135, 136.

102, 103 still has no hits and partially co-elutes with another peak.

[JMB]

Peter,

Thank you for the update, the 60m column certainly pulls the overlap apart.
IIRC, aldehydes can lose that -H to give an (M-1) ion, and that could explain m/z 135.
So the mystery metabolite (MM) possibly does not contain S, but you have a detector for both N and S to hand; the presence/absence of N and S will be important for the HR accurate mass determination.

In the longer term, you may want to consider how to isolate enough of the MM to acquire the 1H NMR spectrum, since that is typically also required for a new compound.

Regards,

JMB

[Peter Apps]

The first run on the PFPD shows no sulfur. Dimethyl sulfoxide shows a decent peak but there is a tiny chance that the S-squared response of the PFPD leaves the mystery compound in the noise. I will re-run it with a more concentrated extract.

[JMB]

Peter,

Negative response for S (if subsequently confirmed) is a good end to the week, because it removes another variable from the chemical composition.

Regards,

JMB

[Peter Apps]

Thanks to Andre de Villiers at Stellenbosch I have the following high res data for the m/z 102/103 mystery peak;

m/z 103.026344: C3H5NO3 (-0.490 ppm)
m/z 102.018536: C3H4NO3 (-0.328 ppm)
m/z 101.010743: C3H3NO3 (-0.014 ppm)
m/z 75.031504: C2H5NO2 (0.322 ppm)
m/z 69.992453: C2NO2 (1.405 ppm)
m/z 59.012889: C2H3O2 (2.258 ppm)
m/z 58.028823: C2H4NO (1.427 ppm)

The molecular ion is too weak for sensible data. We are looking into ESI-MS for that.

Is a structure possible from those data?

Thanks in advance.

[JMB]

Peter,

Here are my initial thoughts on the accurate mass data,

m/z 103 [C3H5NO3] needs 30 amu [CH2O; N2H2; C2H6]] for MW 133

m/z 102 [C3H4NO3] needs 31 amu [-OCH3] for MW 133

m/z 101 [C3H3NO3] needs 32 amu [CH3OH; O2] for MW 133

As a first pass, this suggests that MW 133 is C4H7NO4

The Royal Society of Chemistry has a very nice website, www.chemspider.com,
that can be searched by elemental composition (and has MS/NMR spectra for some compounds); C4H7NO4 brings up 170 matches, starting with the aspartic acids (monoisotopic mass, 133.037508).

You might want to review these matches for structural appropriateness vs the known metabolites in the urine of dogs etc, and vs the GC characteristics you have in hand.

ESI/MS accurate mass of m/z 133 should be your next step.

Andre will be able to advise on "tricks" (TMS; rxn with 1:1 CH3OH/CD4OD etc) to ID structural features.

Regards,
JMB

[JMB]

Peter,

One thing that I had forgotten this morning is that from the empirical molecular formula, you can calculate the degree of unsaturation,

Rings + Double bonds = C - H/2 +N/2 +1

R + DB = 4 - 7/2 + 1/2 + 1
= 2
e.g. >C=O ; >C = C< ; a ring

Regards,
JMB