Development of High-Performance Chemical Isotope Labeling LC-MS for Profiling the Carbonyl Submetabolome.

Liang Li,Shuang Zhao,Margot Dawe,Kevin Guo
DOI: https://doi.org/10.1021/acs.analchem.7b01098
IF: 7.4
2017-05-25
Analytical Chemistry
Abstract:Metabolites containing a carbonyl group represent several important classes of molecules including various forms of ketones and aldehydes such as steroids and sugars. We report a high-performance chemical isotope labeling (CIL) LC-MS method for profiling the carbonyl submetabolome with high coverage and high accuracy and precision of relative quantification. This method is based on the use of dansylhydrazine (DnsHz) labeling of carbonyl metabolites to change their chemical and physical properties to such an extent that the labeled metabolites can be efficiently separated by reversed phase LC and ionized by electrospray ionization MS. In the analysis of six standards representing different carbonyl classes, acetaldehyde could be ionized only after labeling and MS signals were significantly increased for other 5 standards with an enhancement factor ranging from ∼15-fold for androsterone to ∼940-fold for 2-butanone. Differential 12C- and 13C-DnsHz labeling was developed for quantifying metabolic differences in comparative samples where individual samples were separately labeled with 12C-labeling and spiked with a 13C-labeled pooled sample, followed by LC-MS analysis, peak pair picking, and peak intensity ratio measurement. In the replicate analysis of a 1:1 12C-/13C-labeled human urine mixture (n = 6), an average of 2030 ± 39 pairs per run were detected with 1737 pairs in common, indicating the possibility of detecting a large number of carbonyl metabolites as well as high reproducibility of peak pair detection. The average RSD of the peak pair ratios was 7.6%, and 95.6% of the pairs had a RSD value of less than 20%, demonstrating high precision for peak ratio measurement. In addition, the ratios of most peak pairs were close to the expected value of 1.0 (e.g., 95.5% of them had ratios of between 0.67 and 1.5), showing the high accuracy of the method. For metabolite identification, a library of DnsHz-labeled standards was constructed, including 78 carbonyl metabolites with each containing MS, retention time (RT), and MS/MS information. This library and an online search program for labeled carbonyl metabolite identification based on MS, RT, and MS/MS matches have been implemented in a freely available Website, www.mycompoundid.org . Using this library, out of the 1737 peak pairs detected in urine, 33 metabolites were positively identified. In addition, 1333 peak pairs could be matched to the metabolome databases with most of them belonging to the carbonyl metabolites. These results show that 12C-/13C-DnsHz labeling LC-MS is a useful tool for profiling the carbonyl submetabolome of complex samples with high coverage.
Chemistry,Medicine
What problem does this paper attempt to address?