Post on 11-Feb-2021
Identifying Unknowns: A challenge for Metabolomics
Jeevan K. Prasain, PhDDepartment of Pharmacology and
Toxicology, UABjprasain@uab.edu
Outline• Introduction• How to interpret LC-MS and MS/MS
data.• Identification of some conjugated
metabolites.• Conclusions
Introduction• Identification of metabolites (lipids or any other
metabolites) at a molecular level presents a great challenge due to their structural diversity (isobars and isomers) and dynamic metabolism.
• Considering the number of metabolites is >2000,000, there is a lack of commercial analytical standards (only a few thousands available) or comprehensive databases.– Note that there is the opportunity to make specific metabolite
standards through the NIH Common Fund– Go to http://metabolomicsworkbench.org
• Inclusion of many artifacts in database.• Structural complexity of metabolites.• Low concentrations and difficult to isolate.
Metabolome
LC-MS
NMRGC-MS
CE-MSPDA
Adapted from Moco et al., Trend in Analytical Chemistry, 2007
Majority of metabolites are yet to be identified; LC-MSand NMR are the most commonly used analytical techniques
for metabolite identification
Keys to identifying chemical structures by mass spectrometry
• Combination of the following:
o Retention time in LCo Accurate masso Isotope distributiono MS/MS product ions of a precursor ion
Raw LC-MS data
Interpretation of MS/MS of a precursor ion(accurate mass, isotope data and nitrogen rule)
Putative identification(molecular formula determination)
De novo structure elucidation (NMR and or LC-MS/MS)
Metabolite identification workflow
Metabolomics/chemical database search(known/new or novel)
Data file processed by algorithms(e.g., XCMS)
Platform to process untargeted metabolomic data
• XCMS (developed by the Siuzdak Lab at the Scripps Research Institute) Online, is a web-based version that allows users to easily upload and process LC-MS data. It provides links to METLIN database and produces a list of potential metabolites.
• METLIN (developed by the Siuzdak Lab.) is a metabolite database for metabolomics containing over 64,000 structures and it also has comprehensive tandem mass spectrometry data on over 10,000 molecules.
LCMS-based metabolomics• Detection of intact molecular ions [M+H]+/[M-H]- is
possible with soft ionization such as ESI• High mass accuracy of many instruments (
Points to be considered in LC-MS analysis
• Choice of ionization mode- ESI Vs APCI +ve/-vemodes
• Choice of eluting solvent- methanol Vs acetonitrile• Additives/pH in mobile phase• Molecular ion recognition (adduct formation) • Chromatographic separation- stationary phase C8,
C18 ..• Evaluation of spectral quality- what to look for in a
good quality spectra
Nielsen et al., J Nat Prod. 2011
Adduct formation might complicate metabolite identification
Use of isotope pattern in identification of metabolites
• Very close in mass, but different in isotope patterns.
• Isotope ratio outlier analysis (IROA)– Used for LC‐MS (and possibly GC‐MS)– Designed to distinguish between metabolites of interest and background signals
– Requires uniform labeling at the 95% and 5% 13C‐enrichment levels
Pairing the 5% and 95% 13C‐labeling distinguishes artifactual molecules
Courtesy of IROA Technologies
Value of knowing the carbon #
IROA Technologies
Isotopic pattern intensity of [M-H]- and 13C and [M-H+2]-signals indicates the number of carbons and hetero atoms
in the molecular ion
287
289288
1H = 99.9%, 2H = 0.02%12C = 98.9%, 13C = 1.1%35Cl = 68.1%, 37Cl = 31.9%
50 100 150 200 250m/z
100
75
50
25
035 91
139
151
167
195
223251
287
36 Da-HClCl-
50 100 150 200 250m/z
35 59 133139
151
167
195
223 251 287
3’-chlorodaidzein standard
Daidzein + PMA induced HL-60 cells
Comparison of product ions between known and unknown can help elucidate the unknown
Prasain et al., 2003; Boersma et al., 2003
Good chromatographic separation and accurate massindicate a number of diastereoisomer/enantiomer
of PGF2alpha in worm extracts
m/z, amu40 100 160 220 280 340
3.8e6193.1
309.2164.9
171.2 291.1208.9
247.2191.1111.2 353.4263.1255.0173.0
229.2273.2
137.1 181.2281.2
219.043.2 113.0
124.8 299.2
335.159.3 138.8
ESI-MS/MS of the [M-H]- from PGF2a m/z 353 using a quadrupole mass spectrometer
O-
O
OH
HO
HO
Fragmentation scheme of PGF2a [M-H]- m/z 353
Ions m/z 309, 291, 273 and 193 are indicative of F2‐ring
Only in chiral normal phase column,PGF2alpha and its enantiomer can be distinguished
McKnight et al., (Science, 2014)
Great similarities between C. elegans and Cox dKO pups PGF2 profile
Great similarities between C. elegans and Cox dKO pups PGF2 profile
220 280 300 320 360 380 m/z0
%
0
100
%
255.050
256.057
297.043
267.037
268.041281.051
307.065321.046
363.046335.061351.044 381.055
100
240 260 340
[A]
[B]
-162 Da
Yo+
-120 Da
O
OH
O
O
O
OHOH
CH2OH
OH
daidzin
Puerarin
O
O OH
HO
O
HO
OH
OH
CH2OH
Ions break at weakest points; difference in O- and C-glucoside fragmentation
Prasain et al., J. Agric. Food Chem., 2003
m/z
100 311
Identification of unknowns: comparing the MS/MS spectra of the known compound (puerarin)
150 200 250 300 350 400
%
0
255268
282
‐120 431
O
OOH
HO
O
HO
OH
OH
H2COH
+OH
200 250 300 350 400
100
50
0
267 295
\
415
‐120
Monohydroxylatedpuerarin
Puerarin
Mass/Charge120 180 240 300 360 420 480 540 600 660 720 780 8400
26000 283.2637
419.2557
721.4807
437.2651301.2168152.9960
722.4945284.2669
420.2642808.5115
281.2476438.2742302.2204257.2272 723.4921455.2236
-87 Da
Nitrogen rule-Odd number of nitrogens = odd MW
Even nitrogens = even MW
120 180 240 300 360 420 480 540 600 660 720 780 840 9000
20000 883.5366
884.5405
301.2178
283.2648885.5465241.0129
419.2565 810.5110581.3109257.2312302.2218 811.5135223.0022 582.3122526.2935152.996
4420.2590303.2327 810.5906
-302 [M-FA C20:5]-
Accurate mass (
Library search for eicosanoid http://www.lipidmaps.org/
Targeted metabolomics: identification of conjugated
metabolites (glucuronide/sulfate conjugates/glutathione-GSH)
100 200 300 400
m/z
100
50
0
Relativ
e Intensity
(%)
5985
113133 181 224
269
Genistein[M‐H]‐
-176 Da
The loss of anhydrous glucuronic acid (176 Da)-the characteristic of all glucuronides
60 120 180 240 300 360 420 480 540 600 660 720 780 840Mass/Charge, Da
0
70 317.2107
306.0739
272.0877143.0447128.0363 254.0739
821.3264514.2514160.0111210.0907179.0504 249.0358
177.0376288.0615
197.055758.029599.0577
74.0250
-307.07 Da
[M-H]-
Inte
nsity
GSH conjugates can be identified with characteristicfragment ions m/z 306, 272, 254, 210, 179, 160
and 143 in -ve ion mode
HS
HN NH
-O
O
OO
NH2HO
O
m/z 306.0765
Jackson et al. unpublished results
The loss of 80 Da from the parent ion and the presence of m/z 80 in the product ion spectra indicates aromatic sulfate
conjugates
20 60 100 140 180 220 260 300
5.5e6
Intensity, cps
116.9
253.0
134.9
252.0132.991.0224.0
207.980.0
96.9 225.0197.0160.0
[A] OHOO
O
SO
O O-
m/z, amu
20 60 100 140 180 220 260 300 340
1.4e7
Intensity, cps
121.0
119.0
135.0
79.993.0 241.0
147.091.0 320.9
[B]
OO
OH
SO
O
O-
80 Da
80 Da
Aliphatic sulfates typically show m/z 97 (HSO4-), whereas aromatics show m/z 80 (SO3-)
Weidolf et al. Biomed. and Environ. Mass Spec. 1988
Ionization mode Scan
Glucuronides pos/neg NL 176 amuHexose sugar pos/neg NL 162 amuPentose sugar pos/neg NL 132 amuPhenolic sulphate pos NL 80 amuPhosphate neg Prec m/z 79, NL 80/98 GSH Pos/neg NL 129 amuGSH Negtaurines Pos Precursor of m/z 126N-acetylcysteins neg NL 129 amu
NL = neutral loss.
Conjugate
Characteristic fragmentation of metabolite conjugates by MS/MS
Kostiainen et al., 2003
Prec m/z 272
Conclusions• Identifying uncharacterized metabolites in biological
samples is a significant analytical challenge and it requires integrated analytical approaches.
• A well separated metabolite signals in LC-MS/MS is crucial for unambiguous identification of new chemical entity.
• Data processing and data analysis are important for putative identifications.
• The use of high-resolution MS and MSn provides important information to propose structures of fragment and molecular ions.
• Combination of MS with NMR (eg. LC-NMR) is one of the most powerful analytical strategies for structure elucidation of novel metabolites.
Acknowledgements• Stephen Barnes, PhD (UAB)• Michael Miller, PhD (UAB)• Lalita A. Shevde, PhD (UAB) • Ray Moore (ex-TMPL)• Landon Wilson (TMPL)• William P. Jackson (UAB)