Measuring Biomarkers in Blood

Developing a Robust, High Throughput LC-MS Method

  • Fig. 1 Structures of leucine, isoleucine, valine and TMAOFig. 1 Structures of leucine, isoleucine, valine and TMAO
  • Fig. 1 Structures of leucine, isoleucine, valine and TMAO
  • Fig. 2 Separation of the leucine, isoleucine, valine and TMAO using (a) Intrada Amino acid column and (b) ACE C18 –PFP column. Reprinted from Springer: Analytical and Bioanalytical Chemistry [6]
  • Fig. 3 Sample preparation and analysis workflow. Reprinted with some amendments from Reprinted from Springer: Analytical and Bioanalytical Chemistry [6]
  • Fig. 4 Pentafluorophenyl group

The measurement of trimethylamine-N-oxide (TMAO) and branched-chain amino acids (BCAA) in blood is of interest as they are potential indicators of cardiovascular disease risk and diabetes respectively [1-5]. To date, the determination of TMAO and the BCAA is completed as two separate assays, so to improve efficiency the team set out to develop a single LC-MS based method that involved minimal sample preparation. Here a published method is summarised, where an effective column to separate the analytes of interest and also a column with a different mechanism of separation in order to validate the method was identified [6].

Separation and detection of BCAA and TMAO

TMAO is a small polar, zwitterionic, molecule. Due to its polarity, separation is usually achieved on a hydrophobic interaction liquid chromatography (HILIC) column.  There are several phases available for HILIC separations, the most common being silica, zwitterionic and amide. The separation mechanism is complex and depends on both the stationary phase and the analyte, but can include electrostatic, hydrogen bonding and hydrophilic interactions. Having access to different phases also has the advantage of introducing more selectivity and hence separation.  Both silica [7] and amide [8] phases have been used as the stationary phase for separation of TMAO.

There have been numerous reports of the separation of the BCAA; leucine, isoleucine and valine [9-14]. In most instances the separation of the isomers, leucine and isoleucine, prior to detection is a key consideration. Furthermore, the polar nature of the amino acids limits their retention on non-polar columns making separation of these early eluting isomers difficult. Therefore, ion-pairing reagents such as trifluoroacetic acid or heptafluorobutyric acid or chemical derivitising reagents such as dansyl chloride or o-phthalaldehyde are often employed to make the analytes less polar so that they are better retained on non-polar columns.  When sensitive detection of the amino acids is necessary, especially when using UV detection, chemical derivitisation has the added advantage of introducing a UV absorbing chromophore and improving detection limits.

Separation of the native amino acids, when using sensitive MS detection, is preferred, but the need to resolve leucine and isoleucine remains. MS detects the molecular ion (m/z) and being isomers leucine and isoleucine have the same m/z ion (132/+1), after protonation). Furthermore, as their structures differ only in the location of the methyl substitution on the hydrocarbon chain they also share key fragment ions 132→41 and 132→86 and to a lesser extent the fragment ions 132→69 and 132→43 (Fig. 1). In tandem mass spectrometry, usually the most intense of these ions (132→86) is used for quantification purposes. The separation of the native BCAA has been achieved using both HILIC [12,13] and reversed [14] phase separation mechanisms.

Choosing a separation column

Seven separation columns were tested: two C18 columns (ACE C18, Waters Acquity BEH C18), one PFP column (Hypersil Gold Thermo PFP), one C18-PFP (ACE C18-PFP); one amino acid specific column (Intrada Amino Acid); and two HILIC (Waters BEH Amide and Thermo Syncronis) for their ability to resolve the four analytes, and in particular leucine and isoleucine. While there was no attempt made to optimise the mobile phase composition for each column, conditions that would likely promote separation of the analytes were adopted. For example, for the reversed phase columns (C18 and PFP) a highly aqueous mobile phase consisting of 99% acidified water was held for two minutes to maximise retention of the amino acids before increasing the organic content to 10 % over 2 minutes. The flow rate was low (0.2 ml/min), again to aid separation. For the HILIC columns, buffered mobile phases (ammonium formate/formic acid at pH 3 and ammonium acetate/acetic acid at pH 5) were adopted to ensure repeatable retention times, a weakness of HILIC separations.

The two C18 columns achieved baseline resolution of the isomers (R = 1.65, Waters Acquity BEH C18; R= 2.05, ACE C18). The resolution equation tells us that a longer column or use of smaller particle size column packing material will result in improved separation, but the trade-off is higher back pressure and longer run times. A more effective way to increase resolution is to alter selectivity. Using a C18 column with embedded pentafluorophenyl (PFP) groups (ACE C18-PFP) improved the separation of leucine and isoleucine (R= 3.07: Fig. 2). The PFP groups (Fig. 3) introduce additional retention mechanisms including pi-pi interactions, dipole-dipole and hydrogen bonding, and enhanced shape selectivity [15]. While both leucine and isoleucine can engage in hydrogen bonding interactions with the electronegative fluorine atoms, shape selectivity is likely to be responsible for the separation of the isomers. The fluorine atoms add rigidity to the phase and can impact the way the analytes interact with the PFP group and the C18 group. Interestingly, employing just a PFP column (Hypersil Gold Thermo PFP) provided inadequate resolution (R = 1.07) of the isomers, indicating that the hydrophobic interactions due to the C18 phase in the ACE C18-PFP column were important in helping resolve leucine and isoleucine. 

TMAO was poorly retained and eluted first on all the reversed phase columns. The columns with a PFP phase were more effective at retaining the TMAO, and this is likely due to the interaction of the negatively charged fluorine atoms on the PFP ring with the positively charged TMAO.

The HILIC columns were not as effective at separating leucine and isoleucine (R <2).  The best of the columns tested was the Waters BEH amide column at pH 5 which did baseline resolve leucine and isoleucine (R = 1.65). The HILIC columns effectively retained TMAO which eluted after the amino acids. The Intrada Amino Acid column, which is marketed as being able to resolve amino acid isomers [16], did baseline resolve leucine and isoleucine (R = 1.5). TMAO was also well retained.

It was determined that the ACE C18-PFP column best met the need for both full resolution of the amino acid isomers and adequate retention of TMAO.

Mass Spectrometry detection

A Thermo Scientific TSQ Quantitiva Triple Quadrupole Mass Spectrometer was used for the detection of the analytes. It was operated in positive mode and the analytes were ionised by electrospray and monitored in selected reaction monitoring mode. The optimal ion transitions and collision energies were experimentally determined and were confirmed by previous literature. Full details are available in the published article [6].

Sample extraction

The extraction method was kept as simple as possible to minimise analyte loss and contamination, and to allow for large sample throughput. The samples were deproteinated to protect the separation column and MS source by the addition of ice cold acetonitrile (ratio of 1:3 serum:acetonitrile). The acetonitrile was spiked with internal standards (d8valine, d9TMAO, d3leucine) to measure for sample loss and/or matrix effects. The extracted analyte solution was dried and the residue reconstituted in the starting mobile phase conditions. Our current column has separated in excess of 500 plasma samples with no column degradation and excellent retention time repeatability has been maintained (<0.5 % for all analytes). The entire workflow is summarised in Figure 4.

Validation of the method

As we were using the method for a number of studies including: investigating the health impacts of Palaeolithic diet [17], investigating the impact of a fruit diet as part of a cardiovascular health study and more recently a dietary intervention study involving red meat, we validated the method. As part of the validation we wanted to quantify the analytes in real samples using an orthogonal method. An orthogonal method will often involve using a different instrument entirely such as gas chromatography. However, we chose to use LC and the Intrada Amino Acid column, as it clearly used a different mechanism of separation, with the order of elution reversed Leucine<Isoleucine<Valine<TMAO (Fig. 2). Excellent linearity in the concentration range typically found in serum/plasma was demonstrated. We confirmed that the matrix effects were negligible, and used labelled internal standards to correct for any analyte loss during the extraction.

The limits of detection (LODs) and limits of quantitation (LOQs) were determined to ensure the method was appropriate for plasma/serum samples. The LODs (s/n = 3)  and LOQs (s/n = 10) for TMAO were 1 and 6 ng/mL, for leucine and isoleucine were 4 and 8 ng/mL, and for valine were 5 and 15 ng/mL, respectively. These detection limits were typically 10 and 25 times lower than the levels typically reported in serum for BCAA and TMAO respectively [7,14]. 

A recovery experiment (where serum samples were spiked with low med and high concentrations of the analytes) ranged from 97-113 % for the four analytes over the three concentration ranges tested.

Application of method to serum samples

Serum samples (10) were prepared and separated using both the ACE C18-PFP column and the Intrade Amino acid column. There is very good agreement between the methods: for valine and leucine agreement was within 4 %, for TMAO agreement was within 8 % and for isoleucine agreement was within 10 % [6].

Conclusions and developments

This is a simple method for the simultaneous determination of common dietary biomarkers in serum. It is ideal for high-throughput analysis, with 80 samples plus QC samples, standards and blanks prepared and analysed in a 24hr period. The method can also easily be expanded to include other common biomarkers such as creatinine, betaine and carnitine.


Mary C. Boyce

Associate Professor Mary Boyce

Edith Cowan University
Perth, Western Autralia, Australia

Related Articles!



  1. Tang WHW, Wang Z, Li XS, Fan Y, Li DS, Wu Y, et al. Increased Trimethylamine N-Oxide Portends High Mortality Risk Independent of Glycemic Control in Patients with Type 2 Diabetes Mellitus. Clinical Chemistry. 2017;63:297-306.
  2. Tang WHW, Wang Z, Fan Y, Levison B, Hazen JE, Donahue LM, et al. Prognostic Value of Elevated Levels of Intestinal Microbe-Generated Metabolite Trimethylamine-N-Oxide in Patients With Heart Failure: Refining the Gut Hypothesis. Journal of the American College of Cardiology. 2014;64:1908-14.
  3. Wang Z, Klipfell E, Bennett BJ, Koeth R, Levison BS, DuGar B, et al. Gut flora metabolism of phosphatidylcholine promotes cardiovascular disease. Nature. 2011;472:57-63.
  4. Tang WHW, Wang Z, Levison BS, Koeth RA, Britt EB, Fu X, et al. Intestinal Microbial Metabolism of Phosphatidylcholine and Cardiovascular Risk. New England Journal of Medicine. 2013;368:1575-84.
  5. Bergeron N, Williams PT, Lamendella R, Faghihnia N, Grube A, Li X, et al. Diets high in resistant starch increase plasma levels of trimethylamine-N-oxide, a gut microbiome metabolite associated with CVD risk. British Journal of Nutrition. 2016;116:2020-9.
  6. Le, T.T., Shafaei, A., Genoni, A., Christophersen, C., Devine, D., Lo J., Lyons Wall, P., & Boyce, M.C. (2019). Development and validation of a simple LC-MS/MS method for the simultaneous quantitative determination of trimethylamine-N-oxide and branched chain amino acids in human serum. Analytical and Bioanalytical Chemistry, Vol 411, 1019-1028. https://doi:10.1007/s00216-018-1522-8.
  7. Heaney LM, Jones DJL, Mbasu RJ, Ng LL, Suzuki T. High mass accuracy assay for trimethylamine N-oxide using stable-isotope dilution with liquid chromatography coupled to orthogonal acceleration time of flight mass spectrometry with multiple reaction monitoring. Analytical and Bioanalytical Chemistry. 2016;408:797-804.
  8. Ocque AJ, Stubbs JR, Nolin TD. Development and validation of a simple UHPLC–MS/MS method for the simultaneous determination of trimethylamine N-oxide, choline, and betaine in human plasma and urine. Journal of Pharmaceutical and Biomedical Analysis. 2015;109:128-35.
  9. Sowell J, Pollard L, Wood T. Quantification of branched-chain amino acids in blood spots and plasma by liquid chromatography tandem mass spectrometry for the diagnosis of maple syrup urine disease. Journal of Separation Science. 2011;34:631-9.
  10. Moldoveanu SC, Zhu J, Qian N. Free amino acids analysis by liquid chromatography with tandem mass spectrometry in several botanicals with antioxidant character. Journal of Separation Science. 2015;38:2208-22.
  11. Nemkov T, D’Alessandro A, Hansen KC. Three-minute method for amino acid analysis by UHPLC and high-resolution quadrupole orbitrap mass spectrometry. Amino Acids. 2015;47:2345-57.
  12. Guerrasio R, Haberhauer-Troyer C, Mattanovich D, Koellensperger G, Hann S. Metabolic profiling of amino acids in cellular samples via zwitterionic sub-2 μm particle size HILIC-MS/MS and a uniformly 13C labeled internal standard. Analytical and Bioanalytical Chemistry. 2014;406:915-22.
  13. Kato M, Kato H, Eyama S, Takatsu A. Application of amino acid analysis using hydrophilic interaction liquid chromatography coupled with isotope dilution mass spectrometry for peptide and protein quantification. Journal of Chromatography B. 2009;877:3059-64.
  14. Yang R, Dong J, Guo H, Li H, Wang S, Zhao H, et al. Rapid and Precise Measurement of Serum Branched-Chain and Aromatic Amino Acids by Isotope Dilution Liquid Chromatography Tandem Mass Spectrometry. PLOS ONE. 2013;8:e81144.
  15. ACE C18-PFP Technical brochure.
  16. Amino Acids Separation Column for LC-MS. USA: Imtakt USA.
  17. Genoni, A., Lo, J., Lyons-Wall, P., Boyce, M.C., Christophersen, C., Bird, A., & Devine, A. (2019). A Paleolithic diet lowers resistant starch intake but does not affect serum TMAO concentrations in healthy women. British Journal of Nutrition, Vol 121, 322-329. http//doi:10.1017/S000711451800329X.


Edith Cowan University

You may also be interested in

Register now!

The latest information directly via newsletter.

To prevent automated spam submissions leave this field empty.