Skip to Main Content
Skip Nav Destination

A key aim of using a metabolomics approach is to obtain (ultimately) an answer to a well-defined biological question. To this end, state-of-the-art analytical separation techniques are currently used for the reliable profiling of (endogenous) metabolites in a wide range of biological samples. Within metabolomics, capillary electrophoresis–mass spectrometry (CE-MS) has become a very strong analytical tool for the selective profiling of polar and charged metabolites. In CE, compounds are separated according to their charge-to-size ratio and, therefore, the separation mechanism of this approach is fundamentally different from chromatographic-based separation techniques. As such, CE-MS provides complementary information on the metabolic composition of biological samples. In this chapter, the utility of CE and CE-MS for the analysis of (highly) polar and charged metabolites is described from a historical perspective. Attention is devoted to some research works from a few decades ago in which the value of CE for the selective analysis of a few (endogenous) metabolites in human body fluids was demonstrated. Then, seminal research works on the development of CE-MS methods for targeted and non-targeted metabolomics studies are discussed. Finally, the current situation of CE-MS in metabolomics is considered and a view on where this approach may head to is provided.

One of the major goals of using a metabolomics approach is to obtain an answer to a specific biological/clinical question.1  For this purpose, advanced analytical separation techniques are employed for the reliable profiling of (endogenous) metabolites in various biological samples. The most recent version of the Human Metabolome Database (HMDB version 4.0) contains more than 100 000 metabolite entries including both water-soluble and lipid-soluble metabolites, as well as metabolites that would be considered as either abundant (>1 µM) or low-abundant (<1 nM).2  This number of metabolites and their concentration range clearly indicate that various analytical separation techniques with complementary separation mechanisms, like nuclear magnetic resonance (NMR) spectroscopy and mass spectrometry (MS) hyphenated to gas chromatography (GC), liquid chromatography (LC) and/or capillary electrophoresis (CE), are needed in order to analyze a wide range of (endogenous) metabolites in a given biological sample.3  For example, for the characterization of (endogenous) metabolites in human serum, a number of analytical separation techniques with different separation mechanisms have been used, and this strategy allowed the detection of more than 4000 metabolites at concentration levels from the low nM- to the high mM-range.4 

Within the metabolomics field, two analytical strategies can be distinguished, i.e. the non-targeted approach, also known as global metabolic profiling, and the targeted approach.5–8  In non-targeted metabolomics, the aim is to analyze as many (endogenous) metabolites as possible in a given biological sample without having a priori knowledge on the nature and identity of the measured compounds. In targeted metabolomics, the analysis is often focused on the generation of quantitative data for pre-selected metabolites. Both non-targeted and targeted approaches can be employed within a single metabolomics study, where the first approach is generally used for the screening of differential metabolites (or metabolic indicators) for diseases. In the second approach, the statistically relevant differential metabolites may then be quantified and confirmed, preferably employing standardized operating procedures.9  Reproducibility and validation of confirmed differential metabolites for a certain disease require further assessment in large sample cohorts, preferably at multiple sites, in order to determine whether these compounds can be used as biomarkers.10 

At present, the analytical techniques most often used for global metabolic profiling studies are 1H NMR spectroscopy, reversed-phase ultra performance (RP-UP)LC-MS and GC-MS.11–16  Notably, RP-UPLC-MS methods using columns with sub-2 µm porous particles or core–shell silica particles have received a lot of attention for the efficient and fast profiling of metabolites in complex samples.17–19  NMR spectroscopy can still be considered the most robust and reliable analytical tool for high-throughput and reproducible metabolic profiling of body fluids (such as for example serum and urine) using minimal sample pretreatment. Typically, this approach provides quantitative data for metabolites present in the µM to mM concentration range employing a data acquisition time of 5 min per sample.11  However, evaluation of recorded NMR data may be considered as a challenging task due to the complex spectra and superimposition of signals at certain chemical shift regions. GC-MS has a well-known track record for the profiling of (endogenous) metabolites in various body fluids in a clinical setting due to its high separation efficiency and sensitivity.20,21  Also, the reliable identification of compounds by GC-MS is a very strong asset for metabolomics studies. However, this approach is not suitable for the analysis of non-volatile, thermolabile, and/or highly polar compounds and, therefore, derivatization is usually needed to yield volatile and thermostable compounds. Derivatization is generally a time-consuming procedure, thereby potentially limiting the high-throughput capacity of GC-MS for metabolic profiling of large sets of clinical samples.

Both LC-MS and CE-MS can be used for the analysis of polar and non-volatile metabolites in biological samples without using derivatization and laborious sample pretreatment procedures. As indicated above, RP-UPLC-MS has become a key technique for global metabolic profiling studies.22–24  A broad array of metabolites can be analyzed by RP-LC approaches; however, the hydrophobic stationary phases generally used do not provide sufficient retention and selectivity for (highly) polar and charged metabolites. To enable the analysis of such compounds by RP-LC-MS, ion-pairing agents like tributylamine or hexylamine can be added to the mobile phase.25–28  However, the use of ion-pair agents in RPLC-MS may result in severe ion suppression and it may contaminate the ion source and ion optics. Furthermore, stability and re-equilibration of the column prior to the next injection may also be considered an issue in ion-pair RP-LC-MS.8  An emerging liquid chromatographic separation tool for metabolomics is hydrophilic interaction LC (HILIC), in which a polar stationary phase is used in combination with aqueous organic eluents. This approach is increasingly employed in combination with RP-LC-MS for global metabolic profiling studies.29 

In this chapter, attention will be devoted to the potential of CE and especially CE-MS for the profiling of (endogenous) metabolites in biological samples. In comparison to chromatographic-based separation techniques, CE-MS is not as often applied in bioanalysis and metabolomics. However, this approach has some unique analytical characteristics for metabolomics studies.30,31  In CE, referring in this context to capillary zone electrophoresis (CZE), compounds are separated on the basis of differences in their charge-to-size ratio. Therefore, the separation mechanism of CE is fundamentally different from that of chromatographic-based separation techniques. An important feature of CE is the intrinsically high separation efficiency as a result of the flat flow profile of the electro-osmotic flow (EOF), making this technique well-suited for the selective analysis of polar and charged metabolites in biological samples. Moreover, the main source contributing to band broadening is longitudinal diffusion as there is no mass transfer in CE separations. As only small sample volumes are needed (∼10–50 nL), CE is highly suited for the analysis of size-limited biological samples, thereby enabling volume-restricted metabolomics studies.

Concerning the analysis of polar and charged metabolites, both CE and HILIC may be considered as useful analytical tools for this purpose. However, small injection volumes are typically used in CE, resulting in compromised concentration sensitivities for global metabolic profiling studies by CE-MS. The concentration sensitivity of CE can be improved by the use of electrokinetic- or chromatographic-based preconcentration techniques, as outlined in Chapters 5 and 6, respectively. Though both CE and HILIC are useful analytical separation techniques for the analysis of polar and ionic metabolites, recent studies have indicated a significant degree of orthogonality between HILIC and CE for metabolomics.32,33  A comparison of CE-MS with other analytical separation techniques for metabolomics is provided in Chapter 8.

Here, CE and CE-MS approaches developed for metabolite analysis and global metabolic profiling are discussed. The chapter starts by providing a summary of the various CE separation modes used in CE and CE-MS for metabolomics. Special attention will be devoted to research work that eventually resulted in the development of CE-MS systems for global metabolic profiling of biological samples. Finally, the current situation of CE-MS in metabolomics is considered and some thoughts on where this approach may head to is provided.

CE can be employed in a number of separation modes, such as CZE, normally referred to as “CE” and also in this chapter, capillary electrochromatography (CEC), micellar electrokinetic chromatography (MEKC), capillary gel electrophoresis (CGE), capillary isotachophoresis (ITP), capillary iso-electric focusing (cIEF) and microemulsion electrokinetic chromatography (MEEKC). In the main CE separation mode, compounds are separated according to differences in their intrinsic electrophoretic mobility, which is mainly dependent on the size (radius) and charge of the compound. Moreover, this separation mode is also compatible with ESI-MS as volatile buffers can be used for the analysis. The CE separation mode has been used for decades now for the profiling of (endogenous) metabolites in various biological samples. Jellum et al. used CE for the profiling of organic acids in human body fluids in order to screen various human metabolic diseases.34–36  It was found that human urine samples could be injected directly into the CE instrument without employing any sample pretreatment.35  More than 50 metabolites were analyzed in 15 min and identification of a selected number of metabolites was based on migration times and characteristic diode-array spectra. This approach was used for the analysis of urine from patients with metabolic diseases, such as for the screening of pyroglutamic aciduria (Figure 1.1). This work can be considered as one of the first CE-based clinical metabolomics studies, in which the potential of CE for the analysis of small biological samples in biomedicine and clinical diagnosis was already envisioned.

Figure 1.1

CE diode-array analysis of urine from a patient with pyroglutamic aciduria (bottom) and a control subject (top). Hydrodynamic injection was used to inject the urine (7 nL) without any sample pretreatment. Borate buffer (300 mM, pH 8.5) was used, voltage 30 kV. Capillary: effective length 56 cm, 50 µm I.D. The electropherograms were corrected for different degrees of urinary dilution as determined by the creatinine values. Reproduced from J. Chromatogr. B: Biomed. Appl., 683, E. Jellum, H. Dollekamp and C. Blessum, Capillary electrophoresis for clinical problem solving: analysis of urinary diagnostic metabolites and serum proteins, 55–65.35  Copyright 1996, with permission from Elsevier.

Figure 1.1

CE diode-array analysis of urine from a patient with pyroglutamic aciduria (bottom) and a control subject (top). Hydrodynamic injection was used to inject the urine (7 nL) without any sample pretreatment. Borate buffer (300 mM, pH 8.5) was used, voltage 30 kV. Capillary: effective length 56 cm, 50 µm I.D. The electropherograms were corrected for different degrees of urinary dilution as determined by the creatinine values. Reproduced from J. Chromatogr. B: Biomed. Appl., 683, E. Jellum, H. Dollekamp and C. Blessum, Capillary electrophoresis for clinical problem solving: analysis of urinary diagnostic metabolites and serum proteins, 55–65.35  Copyright 1996, with permission from Elsevier.

Close modal

Another pioneering work in the development of CE methods for the screening of metabolic disorders in body fluids emerged from Coral Barbas and co-workers.37–39  This group developed a relatively fast and automated CE method employing UV absorbance detection for the analysis of 27 organic acids in human urine.39  With this method, organic acids in urine samples could be measured directly after centrifugation and dilution. Electrophoretic separation was performed in reversed polarity mode employing a polyacrylamide-coated capillary. Phosphate buffer (pH 6.0) containing a small amount of methanol was used as a background electrolyte (BGE). The developed CE method allowed the reliable determination of methylmalonic, pyroglutamic, and glutaric acids in urine samples of patients with diseases related to these compounds. Also, other groups have developed CE methods for the analysis of organic acids in a clinical context.40–44  For example, Dolnik et al. developed a CE method using a capillary coated with linear polyacrylamide and indirect UV detection for the determination of organic acids in human serum.40  Serum could be analyzed directly, that is without using a deproteinization step. Organic acids, such as pyruvate, phosphate, citrate, malate, acetoacetate and lactate could be analyzed in 12 min by this approach. The CE method was used for the determination of organic acids in serum from a paediatric patient with respiratory insufficiency (Figure 1.2).

Figure 1.2

Determination of organic acids in serum from a paediatric patient with respiratory insufficiency. Experimental conditions: linear polyacrylamide-coated capillary, 75 µm I.D., 360 µm O.D., total length 51 cm, effective length 41 cm; BGE, 10 mM e-aminocaproic acid–10 mM mandelic acid; voltage, −20 kV; detection, absorption at 220 nm. Undiluted serum injected by pressure using 100 mbar for 6 s. Reproduced from J. Chromatogr. A, 716, V. Dolnik and J. Dolnikova, Capillary zone electrophoresis of organic acids in serum of critically ill children, 269–277.40  Copyright 1995, with permission from Elsevier.

Figure 1.2

Determination of organic acids in serum from a paediatric patient with respiratory insufficiency. Experimental conditions: linear polyacrylamide-coated capillary, 75 µm I.D., 360 µm O.D., total length 51 cm, effective length 41 cm; BGE, 10 mM e-aminocaproic acid–10 mM mandelic acid; voltage, −20 kV; detection, absorption at 220 nm. Undiluted serum injected by pressure using 100 mbar for 6 s. Reproduced from J. Chromatogr. A, 716, V. Dolnik and J. Dolnikova, Capillary zone electrophoresis of organic acids in serum of critically ill children, 269–277.40  Copyright 1995, with permission from Elsevier.

Close modal

In order to improve selectivity, MEKC, which employs micelles as a pseudo-stationary phase in the BGE, has also been used for the analysis of (endogenous) metabolites in biological samples.45–48  To enhance the metabolic coverage, both CE (referring to CZE) and MEKC were used for metabolic profiling of urine samples from diabetic rats.47  CE employing a polyacrylamide-coated capillary and reverse polarity was used for the profiling of anionic metabolites, whereas MEKC was used for the profiling of neutral and cationic metabolites. As MEKC offers extra selectivity, an extended metabolic profile of the urine sample was obtained by combining the data obtained with both CE methods. An MEKC approach was also developed for the selective analysis of nucleosides in urine samples from both healthy subjects and cancer patients.49  Phenylboronic acid-based SPE columns have been used for the selective preconcentration of nucleosides from human urine and subsequent nucleoside profiling by MEKC. Figure 1.3 shows typical profiles of urinary nucleosides from a uterine cervical cancer patient, a myoma patient and a healthy subject, obtained by MEKC. A total of 15 nucleosides were positively identified in urine samples from eight uterine myoma patients (benign tumor group), 10 uterine cervical cancer patients (malignant tumor group), and 10 healthy females (normal group). A sulfated β-cyclodextrin-modified MEKC method has also been used for metabolic profiling of human urine, allowing the separation of 80 compounds within 25 min.46  Although improved separation and extra selectivity can be obtained using MEKC for metabolic profiling studies, the coupling of MEKC to MS is rather problematic and often provides limited sensitivity.

Figure 1.3

Electropherograms of urinary nucleosides from a uterine cervical cancer patient (malignant), a myoma patient (benign) and a healthy subject (normal). Experimental conditions: electrophoretic separation at 20 kV on a fused-silica capillary (57 cm × 50 µm I.D.) at 20 °C using a BGE of 200 mM SDS–25 mM borate–42.5 mM phosphate (pH 6.7) using hydrodynamic injection (4 seconds at 35 mbar), with simultaneous detection at 210 and 260 nm in dual-channel mode. Peaks: 1 = pseudouridine; 2 = dihydrouridine; 3 = uridine; 4 = cytidine; 5 = 5-methyluridine; 6 = 3-methyluridine; 7 = inosine; 8 = N4-acetylcytidine, 9 = guanosine; 10 = 1-methylguanosine; 11 = adenosine; 12 = xanthosine; 13 = N2-methylguanosine; 14 = N2,N2-dimethylguanosine; 15 = N6-methyladenosine; I.S. = 3-deazauridine. Reproduced from J. Chromatogr. B: Biomed. Sci. Appl., 754, K. R. Kim, S. La, A. Kim, J. H. Kim and H. M. Liebich, Capillary electrophoretic profiling and pattern recognition analysis of urinary nucleosides from uterine myoma and cervical cancer patients, 97–106.49  Copyright 2001, with permission from Elsevier.

Figure 1.3

Electropherograms of urinary nucleosides from a uterine cervical cancer patient (malignant), a myoma patient (benign) and a healthy subject (normal). Experimental conditions: electrophoretic separation at 20 kV on a fused-silica capillary (57 cm × 50 µm I.D.) at 20 °C using a BGE of 200 mM SDS–25 mM borate–42.5 mM phosphate (pH 6.7) using hydrodynamic injection (4 seconds at 35 mbar), with simultaneous detection at 210 and 260 nm in dual-channel mode. Peaks: 1 = pseudouridine; 2 = dihydrouridine; 3 = uridine; 4 = cytidine; 5 = 5-methyluridine; 6 = 3-methyluridine; 7 = inosine; 8 = N4-acetylcytidine, 9 = guanosine; 10 = 1-methylguanosine; 11 = adenosine; 12 = xanthosine; 13 = N2-methylguanosine; 14 = N2,N2-dimethylguanosine; 15 = N6-methyladenosine; I.S. = 3-deazauridine. Reproduced from J. Chromatogr. B: Biomed. Sci. Appl., 754, K. R. Kim, S. La, A. Kim, J. H. Kim and H. M. Liebich, Capillary electrophoretic profiling and pattern recognition analysis of urinary nucleosides from uterine myoma and cervical cancer patients, 97–106.49  Copyright 2001, with permission from Elsevier.

Close modal

The utility of CEC, an analytical separation technique that combines the high peak efficiency of CE with the stationary phase selectivity of LC, has also been assessed for metabolic profiling studies.50,51  For example, a pressure-assisted CEC method was developed for the profiling of metabolites in rat urine and the performance of this approach was compared with chromatographic methods.51 Figure 1.4 clearly shows that CEC was able to measure more urinary metabolites with good resolution as compared to a standard RP-LC or capillary RP-LC method. As UV absorbance was employed for detection, only eight compounds could be identified in the urine sample using comparison with standards. Though CEC can be used for the separation of a wide range of compounds and, as such, is very promising for global metabolic profiling studies, the preparation of CEC columns in a reproducible manner remains a challenging procedure.

Figure 1.4

Comparison of capillary RP-LC (A), RP-LC (B), and pCEC (C) for metabolic analysis of rat urine. RP-LC conditions: column, 4.6 mm × 250 mm, packed with C18, 5 µm particle size; capillary RP-LC conditions: column, 150 µm id, 50 cm total length, packed with C18, 5 µm particle size. Gradient of water (0.02% TFA) and 95% methanol (0.02% TFA) used for LC separation. CEC conditions: column, 150 µm i.d., 50 cm total length, packed with C18, 5 µm particle size; gradient of water (0.01% TFA) and acetonitrile (0.01% TFA) used for separation. Injection volume, 5 µL; detection, UV at 214 nm. Reproduced with permission from ref. 51. Copyright © 2007 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim.

Figure 1.4

Comparison of capillary RP-LC (A), RP-LC (B), and pCEC (C) for metabolic analysis of rat urine. RP-LC conditions: column, 4.6 mm × 250 mm, packed with C18, 5 µm particle size; capillary RP-LC conditions: column, 150 µm id, 50 cm total length, packed with C18, 5 µm particle size. Gradient of water (0.02% TFA) and 95% methanol (0.02% TFA) used for LC separation. CEC conditions: column, 150 µm i.d., 50 cm total length, packed with C18, 5 µm particle size; gradient of water (0.01% TFA) and acetonitrile (0.01% TFA) used for separation. Injection volume, 5 µL; detection, UV at 214 nm. Reproduced with permission from ref. 51. Copyright © 2007 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim.

Close modal

Concerning metabolite analysis by CE in a clinical setting, detection has been mainly performed by UV absorbance detection, but contactless conductivity detection, and electrochemical and laser-induced fluorescence (LIF) detection have also been used for this purpose.52–60  In particular, the use of CE-LIF can be considered a very powerful approach for the selective and highly sensitive analysis of a number of metabolites in small sample amounts. Moreover, CE-LIF can be used for high-speed electrophoretic separations without compromising the separation efficiency. For example, Harstad et al. developed a high-speed microdialysis–capillary electrophoresis (MD-CE) method for monitoring the uptake/release dynamics of branched chain amino acids in 3T3-L1 cells using LIF detection.61  The separation conditions were optimized to resolve isoleucine, leucine, valine, glutamine, alanine, glutamate and many other amino acids in less than 30 s, as shown in Figure 1.5. The peak widths were around 0.2 s with plate numbers above 150 000. The high-speed MD-CE-LIF method allowed the analysis of 15 samples with 8 replicates per sample in less than 2 hours, whereas the same analysis using a traditional offline CE approach would have taken more than 20 hours.

Figure 1.5

Electropherograms from online MD-CE analyses of a 20 µM standard amino acid solution. (A) Separation buffer: 100 mM borate/20 mM HP-β-CD, pH = 10.5; peaks identified: (1) NBD-OH, (2) leucine, isoleucine, (3) valine, (4) citrulline. (B) Separation buffer: 90 mM borate/35 mM α-CD, pH = 9.8; peaks identified: (1) PEA, arginine, (2) threonine, (3) lysine, (4) isoleucine, (5) leucine, (6) ornithine, (7) methionine, (8) phenylalanine, (9) valine, citrulline, (10) α-ABA, (11) histidine, (12) GABA, (13) glutamine, (14) alanine, (15) β-ABA, (16) threonine, (17) β-alanine, (18) glycine, (19) NBD-OH, (20) taurine, (21) internal standard, (22) glutamate, and (23) aspartate. Reproduced with permission from ref. 61. Copyright (2016) American Chemical Society.

Figure 1.5

Electropherograms from online MD-CE analyses of a 20 µM standard amino acid solution. (A) Separation buffer: 100 mM borate/20 mM HP-β-CD, pH = 10.5; peaks identified: (1) NBD-OH, (2) leucine, isoleucine, (3) valine, (4) citrulline. (B) Separation buffer: 90 mM borate/35 mM α-CD, pH = 9.8; peaks identified: (1) PEA, arginine, (2) threonine, (3) lysine, (4) isoleucine, (5) leucine, (6) ornithine, (7) methionine, (8) phenylalanine, (9) valine, citrulline, (10) α-ABA, (11) histidine, (12) GABA, (13) glutamine, (14) alanine, (15) β-ABA, (16) threonine, (17) β-alanine, (18) glycine, (19) NBD-OH, (20) taurine, (21) internal standard, (22) glutamate, and (23) aspartate. Reproduced with permission from ref. 61. Copyright (2016) American Chemical Society.

Close modal

Concerning the use of CE for comparative metabolic profiling studies, the reproducibility of the migration times and peak areas is a critical aspect.6,62  Variability of the migration time is a particularly important issue when employing bare fused-silica capillaries. Adsorption of matrix components and/or analytes to the inner wall of the fused-silica capillaries may cause an irreproducible electro-osmotic-flow (EOF) and consequently irreproducible migration times. Moreover, when bare fused-silica capillaries are used, the separation efficiencies and repeatability of the peak area may be compromised as a result of adverse analyte–capillary wall interactions. Therefore, it is very important to incorporate, in addition to an effective sample pretreatment procedure, an effective rinsing procedure between runs when using a bare fused-silica capillary for metabolic profiling of (especially protein-rich) biological samples. Obviously, such a rinsing procedure will increase the total analysis time per sample. Another way to address this problem is to modify the inner capillary wall of the fused-silica capillaries with polymers, and this approach is further described in Chapter 3. For more details on the development of CE approaches for metabolite analysis and metabolic profiling studies from a historical perspective, the reader is referred to the following reviews.63–68 

In modern metabolomics, the use of a detector providing both a high selectivity and sensitivity is inevitable for the global profiling of metabolites in complex samples. Although detection in CE is still often based on on-column UV absorbance as it is easy to use and broadly applicable, it lacks sensitivity as a result of the small optical path length due to the capillary inner diameter. Moreover, the selectivity of UV detection is limited as it hardly provides any structural information on the compounds. MS has emerged as a key technology for the selective and sensitive analysis of metabolites in biological samples, providing the ability to quantify and identify metabolites. The combination of MS with a front-end separation technique is often needed in order to reduce ion suppression/enhancement, to allow the separation of isobars and isomers, and to gain additional information on the physico-chemical properties of the metabolites, which may facilitate the identification process. CE-MS-based metabolomics studies generally require the use of a sample preparation step to selectively remove the highly polar and charged metabolites from proteins and lipids in a given biological sample. Sample pretreatment may result in the loss of metabolites and, therefore, proper optimization of this step in the analytical workflow designed for a metabolomics study is key for obtaining reliable metabolic profiles. In Chapter 2, attention is paid to sample preparation strategies for CE-MS-based metabolomics studies.

The coupling of CE to MS can be performed with a sheath-liquid or a sheathless interface. Until now, the sheath-liquid interface has most widely been used for CE-MS-based metabolomics studies.31,69,70  Details about interfacing techniques for CE-MS and their utility for metabolomics can be found in Chapter 4. The ionization technique commonly used in CE-MS for metabolomics is ESI and, therefore, the selection of the separation conditions is very critical as ESI is susceptible to analyte signal suppression by relatively high buffer concentrations, non-volatile components and surfactants. Moreover, non-volatile components may cause contamination of the ion source and ion optics and as a result high background signals. Therefore, volatile BGEs, such as formic acid or ammonium acetate, are often used as BGEs for CE-MS studies.

One of the first applications reporting the use of CE-MS for metabolite analysis in a clinical setting came from the group of Egil Jellum.71  In this work, a CE-MS/MS method was developed for the selective analysis of a set of metabolites in dried blood spots from patients with propionic aciduria, which is a metabolic disorder. CE was coupled to MS via a co-axial sheath-liquid interface and electrophoretic separation was performed with a bare fused-silica capillary using 20 mM ammonium acetate (pH 8.5) as a BGE. Around the same time, Soga et al. developed a CE-MS method for the analysis of 19 naturally occurring amino acids without using any derivatization.72  In order to enable the simultaneous and selective analysis of all the amino acids by CE-MS, a low-pH BGE, i.e. 1 M formic acid (pH 1.8), was used (Figure 1.6). For eight repeated injections, the relative standard deviations for the migration times and peak areas were below 1.2% and 7%, respectively. The CE-MS method provided low µM detection limits for most amino acids (corresponding to femtomole amounts) and its potential was demonstrated by the analysis of soy sauce, beer, sake, and urine. The authors already envisioned at that time that the developed method could potentially be used for the profiling of amino acids and related compounds in body fluids and other biological samples for the screening of disorders in amino acid metabolism. Indeed, soon after the introduction of this method for the selective profiling of amino acids, CE-MS systems for global metabolic profiling studies were introduced.

Figure 1.6

Selected ion electropherograms for a standard mixture of 19 amino acids obtained by CE-MS. Experimental conditions: sample concentration, Cys 125 µmol L−1 and others 250 µmol L−1 each; capillary, fused silica 50 µm i.d. × 100 cm; BGE, 1 M formic acid; applied potential, +30 kV; injection, 3 s at 50 mbar; sheath liquid, 10 µL min−1 of 5 mM ammonium acetate in 50% (v/v) methanol−water. Reproduced with permission from ref. 72. Copyright (2000) American Chemical Society.

Figure 1.6

Selected ion electropherograms for a standard mixture of 19 amino acids obtained by CE-MS. Experimental conditions: sample concentration, Cys 125 µmol L−1 and others 250 µmol L−1 each; capillary, fused silica 50 µm i.d. × 100 cm; BGE, 1 M formic acid; applied potential, +30 kV; injection, 3 s at 50 mbar; sheath liquid, 10 µL min−1 of 5 mM ammonium acetate in 50% (v/v) methanol−water. Reproduced with permission from ref. 72. Copyright (2000) American Chemical Society.

Close modal

The first CE-MS method used for the global profiling of metabolites in biological samples was developed by Soga et al.73  Distinct CE-MS methods were used for the profiling of cationic and anionic metabolites, i.e. the first group was analyzed with a bare fused-silica capillary employing 1 M formic acid (pH 1.8) as the BGE (as initially developed for amino acid profiling), whereas the second group was analyzed with a cationic polymer-coated capillary employing 50 mM ammonium acetate (pH 8.5) as the BGE. A typical result for the analysis of nucleotides and CoA compounds obtained using the latter CE-MS conditions is shown in Figure 1.7. By using both CE-MS approaches, more than 1600 molecular features (a molecular feature is a detected signal that appears in the majority of the measured samples comprising a fixed interval of m/z and migration times) could be observed in an extract from Bacillus subtilis, of which 150 could be identified. The developed CE-MS systems allowed the selective and highly efficient analysis of diverse classes of highly polar metabolites, such as amino acids, nucleosides, small peptides, organic acids, nucleotides and sugar phosphates, which are difficult to analyze by conventional RP-LC-MS methods. More details about the design of CE-MS methods for anionic and cationic metabolic profiling can be found in Chapter 7.

Figure 1.7

Analysis of a standard mixture (100 µM each) of nucleotides and CoA compounds by CE-MS using a cationic-coated capillary. Experimental conditions: BGE, 50 mM ammonium acetate (pH 8.5); separation at −30 kV; injection volume, 30 nL. The numbers in the upper left corner indicate the abundances associated with the tallest peak in the electropherogram, for each m/z. Reproduced with permission from ref. 73. Copyright (2003) American Chemical Society.

Figure 1.7

Analysis of a standard mixture (100 µM each) of nucleotides and CoA compounds by CE-MS using a cationic-coated capillary. Experimental conditions: BGE, 50 mM ammonium acetate (pH 8.5); separation at −30 kV; injection volume, 30 nL. The numbers in the upper left corner indicate the abundances associated with the tallest peak in the electropherogram, for each m/z. Reproduced with permission from ref. 73. Copyright (2003) American Chemical Society.

Close modal

For the coupling of CE to MS, volatile buffers like ammonium acetate, acetic acid or formic acid are often used as a BGE; however, these buffers may not provide optimal electrophoretic separations. In this case, organic modifiers, such as, e.g., isopropanol, methanol or acetonitrile, can be added to the BGE in order to improve the separation and MS detection performance of metabolites. For example, in a CE-MS method developed for the analysis of free amino acids, a BGE of 2 M formic acid (pH 1.8) containing 20% methanol was employed in order to obtain a full baseline separation of the isomers leucine and isoleucine.74 

Until now, most CE-MS-based metabolomics studies have been performed with time-of-flight (TOF) MS instruments as they provide a high resolution and a high mass accuracy; the latter is especially important for metabolite identification purposes.31,69  Moreover, the high spectral acquisition rate of a TOF mass analyzer is fully compatible with the relatively fast and transient signals generated during highly efficient CE separations, i.e., a high number of data points can be collected across a sharp CE peak by a TOF-MS instrument. The current generation of triple quadrupole and ion trap MS instruments are also fully compatible with highly efficient CE analyses. For a complete overview of CE-MS-based metabolomics studies reported over the past decade, the reader is referred to the following reviews.18,64,65,70,75,76 

In metabolomics, CE-MS has emerged as a very useful analytical tool for the profiling of highly polar and charged metabolites in biological samples, providing complementary biochemical information in comparison to chromatographic-based separation techniques. Until now, most CE-MS-based metabolomics studies have been focused on the analysis of polar and charged metabolites, including amino acids, amines, nucleosides, nucleotides, organic acids, sugar phosphates and small peptides, using an aqueous BGE system. These types of metabolites are often poorly retained by standard RP-LC-MS systems. It would be interesting to assess the utility of CE-MS also for non-polar and charged metabolites, such as charged hydrophobic lipids, in order to further expand the metabolic coverage of this approach.

Over the past few years, significant developments have been made in CE-MS methodology for metabolomics studies in order to address important analytical challenges, including sensitivity (or metabolome coverage), sample throughput, strategies for data analysis and unknown metabolite identification. All these aspects are considered in various chapters throughout this book. As metabolic profiling approaches are now more widely employed for large-scale phenotyping studies (for example, biological samples from Biobanks), the ability to measure hundreds of samples with a reliable high-throughput precision has become very important. In this context, the recently developed multi-segment injection (MSI-)CE-MS method for global metabolic profiling can be considered a very pragmatic tool (as outlined in Chapter 12). This approach can be used for serial hydrodynamic injection of seven or more discrete sample segments within a single capillary (Figure 1.8), thereby increasing sample throughput up to one order of magnitude without ion suppression and maintaining separation performance.77 

Figure 1.8

(A) Multiplexed separation based on serial injection of seven discrete sample segments within a single capillary by MSI-CE-MS; (B) ions migrate as a series of zones in free solution prior to ionization; (C) the procedure enables reliable quantification of polar metabolites and their isomers in different samples as ionization occurs within a short-time interval (≈2–6 min) under steady-state conditions when using TOF-MS. Reproduced with permission from ref. 77. Copyright (2013) American Chemical Society.

Figure 1.8

(A) Multiplexed separation based on serial injection of seven discrete sample segments within a single capillary by MSI-CE-MS; (B) ions migrate as a series of zones in free solution prior to ionization; (C) the procedure enables reliable quantification of polar metabolites and their isomers in different samples as ionization occurs within a short-time interval (≈2–6 min) under steady-state conditions when using TOF-MS. Reproduced with permission from ref. 77. Copyright (2013) American Chemical Society.

Close modal

So far, CE-MS has only been used by a limited number of groups for metabolomics with the majority of the studies reported by Soga and co-workers. Performance metrics for large-scale metabolic profiling studies by CE-MS including standard operating procedures and vendor support are urgently needed in order to expand the role of CE-MS in metabolomics and bioanalysis in general. A highly encouraging study in this context emerged very recently,78  in which CE-MS was used for the reliable profiling and absolute quantification of polar and charged metabolites in plasma samples from the Tsuruoka Metabolomics Cohort Study (>10 000 human subjects).

Dr Rawi Ramautar would like to acknowledge the financial support of the Veni and Vidi grant scheme of the Netherlands Organisation for Scientific Research (NWO Veni 722.013.008 and Vidi 723.016.003).

1.
Ramautar
 
R.
Berger
 
R.
van der Greef
 
J.
Hankemeier
 
T.
Curr. Opin. Chem. Biol.
2013
, vol. 
17
 (pg. 
841
-
846
)
2.
Wishart
 
D. S.
Feunang
 
Y. D.
Marcu
 
A.
Guo
 
A. C.
Liang
 
K.
Vazquez-Fresno
 
R.
Sajed
 
T.
Johnson
 
D.
Li
 
C.
Karu
 
N.
Sayeeda
 
Z.
Lo
 
E.
Assempour
 
N.
Berjanskii
 
M.
Singhal
 
S.
Arndt
 
D.
Liang
 
Y.
Badran
 
H.
Grant
 
J.
Serra-Cayuela
 
A.
Liu
 
Y.
Mandal
 
R.
Neveu
 
V.
Pon
 
A.
Knox
 
C.
Wilson
 
M.
Manach
 
C.
Scalbert
 
A.
Nucleic Acids Res.
2018
, vol. 
46
 (pg. 
D608
-
D617
)
3.
Gonzalez-Dominguez
 
A.
Duran-Guerrero
 
E.
Fernandez-Recamales
 
A.
Maria Lechuga-Sancho
 
A.
Sayago
 
A.
Schwarz
 
M.
Segundo
 
C.
Gonzalez-Dominguez
 
R.
Curr. Top. Med. Chem.
2017
, vol. 
17
 (pg. 
3289
-
3295
)
4.
Psychogios
 
N.
Hau
 
D. D.
Peng
 
J.
Guo
 
A. C.
Mandal
 
R.
Bouatra
 
S.
Sinelnikov
 
I.
Krishnamurthy
 
R.
Eisner
 
R.
Gautam
 
B.
Young
 
N.
Xia
 
J.
Knox
 
C.
Dong
 
E.
Huang
 
P.
Hollander
 
Z.
Pedersen
 
T. L.
Smith
 
S. R.
Bamforth
 
F.
Greiner
 
R.
McManus
 
B.
Newman
 
J. W.
Goodfriend
 
T.
Wishart
 
D. S.
PLoS One
2011
, vol. 
6
 pg. 
e16957
 
5.
Begou
 
O.
Gika
 
H. G.
Wilson
 
I. D.
Theodoridis
 
G.
Analyst
2017
, vol. 
142
 (pg. 
3079
-
3100
)
6.
Ramautar
 
R.
Adv. Clin. Chem.
2016
, vol. 
74
 (pg. 
1
-
34
)
7.
Dudzik
 
D.
Barbas-Bernardos
 
C.
Garcia
 
A.
Barbas
 
C.
J. Pharm. Biomed. Anal.
2018
, vol. 
147
 (pg. 
149
-
173
)
8.
Ramautar
 
R.
Adv. Clin. Chem.
2016
, vol. 
74
 (pg. 
1
-
34
)
9.
Naz
 
S.
Vallejo
 
M.
Garcia
 
A.
Barbas
 
C.
J. Chromatogr. A
2014
, vol. 
1353
 (pg. 
99
-
105
)
10.
Zhang
 
A.
Sun
 
H.
Yan
 
G.
Wang
 
P.
Wang
 
X.
BioMed Res. Int.
2015
, vol. 
2015
 pg. 
354671
 
11.
Dona
 
A. C.
Jimenez
 
B.
Schafer
 
H.
Humpfer
 
E.
Spraul
 
M.
Lewis
 
M. R.
Pearce
 
J. T.
Holmes
 
E.
Lindon
 
J. C.
Nicholson
 
J. K.
Anal. Chem.
2014
, vol. 
86
 (pg. 
9887
-
9894
)
12.
Beckonert
 
O.
Coen
 
M.
Keun
 
H. C.
Wang
 
Y.
Ebbels
 
T. M.
Holmes
 
E.
Lindon
 
J. C.
Nicholson
 
J. K.
Nat. Protoc.
2010
, vol. 
5
 (pg. 
1019
-
1032
)
13.
Roux
 
A.
Lison
 
D.
Junot
 
C.
Heilier
 
J. F.
Clin. Biochem.
2011
, vol. 
44
 (pg. 
119
-
135
)
14.
Want
 
E. J.
Wilson
 
I. D.
Gika
 
H.
Theodoridis
 
G.
Plumb
 
R. S.
Shockcor
 
J.
Holmes
 
E.
Nicholson
 
J. K.
Nat. Protoc.
2010
, vol. 
5
 (pg. 
1005
-
1018
)
15.
Chan
 
E. C.
Pasikanti
 
K. K.
Nicholson
 
J. K.
Nat. Protoc.
2011
, vol. 
6
 (pg. 
1483
-
1499
)
16.
Fiehn
 
O.
Curr. Protoc. Mol. Biol.
2016
, vol. 
114
 (pg. 
30 34 31
-
30 34 32
)
17.
Yin
 
P.
Xu
 
G.
J. Chromatogr. A
2014
, vol. 
1374
 (pg. 
1
-
13
)
18.
Kuehnbaum
 
N. L.
Britz-McKibbin
 
P.
Chem. Rev.
2013
, vol. 
113
 (pg. 
2437
-
2468
)
19.
Kohler
 
I.
Giera
 
M.
J. Sep. Sci.
2017
, vol. 
40
 (pg. 
93
-
108
)
20.
Koek
 
M. M.
Jellema
 
R. H.
van der Greef
 
J.
Tas
 
A. C.
Hankemeier
 
T.
Metabolomics
2011
, vol. 
7
 (pg. 
307
-
328
)
21.
Koek
 
M. M.
Muilwijk
 
B.
van der Werf
 
M. J.
Hankemeier
 
T.
Anal. Chem.
2006
, vol. 
78
 (pg. 
1272
-
1281
)
22.
Gika
 
H. G.
Zisi
 
C.
Theodoridis
 
G.
Wilson
 
I. D.
J. Chromatogr. B: Anal. Technol. Biomed. Life Sci.
2016
, vol. 
1008
 (pg. 
15
-
25
)
23.
Gika
 
H. G.
Theodoridis
 
G. A.
Plumb
 
R. S.
Wilson
 
I. D.
J. Pharm. Biomed. Anal.
2014
, vol. 
87
 (pg. 
12
-
25
)
24.
Vorkas
 
P. A.
Isaac
 
G.
Anwar
 
M. A.
Davies
 
A. H.
Want
 
E. J.
Nicholson
 
J. K.
Holmes
 
E.
Anal. Chem.
2015
, vol. 
87
 (pg. 
4184
-
4193
)
25.
Lu
 
W.
Clasquin
 
M. F.
Melamud
 
E.
Amador-Noguez
 
D.
Caudy
 
A. A.
Rabinowitz
 
J. D.
Anal. Chem.
2010
, vol. 
82
 (pg. 
3212
-
3221
)
26.
Seifar
 
R. M.
Ras
 
C.
Deshmukh
 
A. T.
Bekers
 
K. M.
Suarez-Mendez
 
C. A.
da Cruz
 
A. L.
van Gulik
 
W. M.
Heijnen
 
J. J.
J. Chromatogr. A
2013
, vol. 
1311
 (pg. 
115
-
120
)
27.
Seifar
 
R. M.
Ras
 
C.
van Dam
 
J. C.
van Gulik
 
W. M.
Heijnen
 
J. J.
van Winden
 
W. A.
Anal. Biochem.
2009
, vol. 
388
 (pg. 
213
-
219
)
28.
Seifar
 
R. M.
Zhao
 
Z.
van Dam
 
J.
van Winden
 
W.
van Gulik
 
W.
Heijnen
 
J. J.
J. Chromatogr. A
2008
, vol. 
1187
 (pg. 
103
-
110
)
29.
Tang
 
D. Q.
Zou
 
L.
Yin
 
X. X.
Ong
 
C. N.
Mass Spectrom. Rev.
2016
, vol. 
35
 (pg. 
574
-
600
)
30.
Zhang
 
W.
Hankemeier
 
T.
Ramautar
 
R.
Curr. Opin. Biotechnol.
2017
, vol. 
43
 (pg. 
1
-
7
)
31.
Hirayama
 
A.
Wakayama
 
M.
Soga
 
T.
TrAC, Trends Anal. Chem.
2014
, vol. 
61
 (pg. 
215
-
222
)
32.
Kok
 
M. G.
Somsen
 
G. W.
de Jong
 
G. J.
Talanta
2015
, vol. 
132
 (pg. 
1
-
7
)
33.
Naz
 
S.
Calderon
 
A. A.
Garcia
 
A.
Gallafrio
 
J.
Mestre
 
R. T.
Gonzalez
 
E. G.
de Cabo
 
C. M.
Delgado
 
M. C.
Balanza
 
J. A.
Simionato
 
A. V.
Vaeza
 
N. N.
Barbas
 
C.
Ruperez
 
F. J.
Electrophoresis
2015
, vol. 
36
 (pg. 
2303
-
2313
)
34.
Jellum
 
E.
Thorsrud
 
A. K.
Time
 
E.
J. Chromatogr.
1991
, vol. 
559
 (pg. 
455
-
465
)
35.
Jellum
 
E.
Dollekamp
 
H.
Blessum
 
C.
J. Chromatogr. B: Biomed. Sci. Appl.
1996
, vol. 
683
 (pg. 
55
-
65
)
36.
Jellum
 
E.
Dollekamp
 
H.
Brunsvig
 
A.
Gislefoss
 
R.
J. Chromatogr. B: Biomed. Sci. Appl.
1997
, vol. 
689
 (pg. 
155
-
164
)
37.
Barbas
 
C.
Adeva
 
N.
Aguilar
 
R.
Rosillo
 
M.
Rubio
 
T.
Castro
 
M.
Clin. Chem.
1998
, vol. 
44
 (pg. 
1340
-
1342
)
38.
Garcia
 
A.
Barbas
 
C.
Aguilar
 
R.
Castro
 
M.
Clin. Chem.
1998
, vol. 
44
 (pg. 
1905
-
1911
)
39.
Garcia
 
A.
Barbas
 
C.
Clin. Chem. Lab. Med.
2003
, vol. 
41
 (pg. 
755
-
761
)
40.
Dolnik
 
V.
Dolnikova
 
J.
J. Chromatogr. A
1995
, vol. 
716
 (pg. 
269
-
277
)
41.
Ramautar
 
R.
Somsen
 
G. W.
de Jong
 
G. J.
Anal. Bioanal. Chem.
2007
, vol. 
387
 (pg. 
293
-
301
)
42.
Shirao
 
M.
Furuta
 
R.
Suzuki
 
S.
Nakazawa
 
H.
Fujita
 
S.
Maruyama
 
T.
J. Chromatogr. A
1994
, vol. 
680
 (pg. 
247
-
251
)
43.
Petucci
 
C. J.
Kantes
 
H. L.
Strein
 
T. G.
Veening
 
H.
J. Chromatogr. B: Biomed. Sci. Appl.
1995
, vol. 
668
 (pg. 
241
-
251
)
44.
Chen
 
H.
Xu
 
Y.
Van Lente
 
F.
Ip
 
M. P.
J. Chromatogr. B: Biomed. Sci. Appl.
1996
, vol. 
679
 (pg. 
49
-
59
)
45.
Guillo
 
C.
Perrett
 
D.
Hanna-Brown
 
M.
Chromatographia
2004
, vol. 
59
 (pg. 
S157
-
S164
)
46.
Guillo
 
C.
Barlow
 
D.
Perrett
 
D.
Hanna-Brown
 
M.
J. Chromatogr. A
2004
, vol. 
1027
 (pg. 
203
-
212
)
47.
Vallejo
 
M.
Angulo
 
S.
Garcia-Martinez
 
D.
Garcia
 
A.
Barbas
 
C.
J. Chromatogr. A
2008
, vol. 
1187
 (pg. 
267
-
274
)
48.
Barbas
 
C.
Vallejo
 
M.
Garcia
 
A.
Barlow
 
D.
Hanna-Brown
 
M.
J. Pharm. Biomed. Anal.
2008
, vol. 
47
 (pg. 
388
-
398
)
49.
Kim
 
K. R.
La
 
S.
Kim
 
A.
Kim
 
J. H.
Liebich
 
H. M.
J. Chromatogr. B: Biomed. Sci. Appl.
2001
, vol. 
754
 (pg. 
97
-
106
)
50.
Zhang
 
H.
Wang
 
Y.
Gu
 
X.
Zhou
 
J.
Yan
 
C.
Electrophoresis
2011
, vol. 
32
 (pg. 
340
-
347
)
51.
Xie
 
G.
Su
 
M.
Li
 
P.
Gu
 
X.
Yan
 
C.
Qiu
 
Y.
Li
 
H.
Jia
 
W.
Electrophoresis
2007
, vol. 
28
 (pg. 
4459
-
4468
)
52.
Hong
 
J.
Baldwin
 
R. P.
J. Capillary Electrophor.
1997
, vol. 
4
 (pg. 
65
-
71
)
53.
Schneede
 
J.
Ueland
 
P. M.
Anal. Chem.
1995
, vol. 
67
 (pg. 
812
-
819
)
54.
Phillips
 
T. M.
Electrophoresis
2018
, vol. 
39
 (pg. 
126
-
135
)
55.
Lorenzo
 
M. P.
Villasenor
 
A.
Ramamoorthy
 
A.
Garcia
 
A.
Electrophoresis
2013
, vol. 
34
 (pg. 
1701
-
1709
)
56.
Lorenzo
 
M. P.
Navarrete
 
A.
Balderas
 
C.
Garcia
 
A.
J. Pharm. Biomed. Anal.
2013
, vol. 
73
 (pg. 
116
-
124
)
57.
Lacna
 
J.
Foret
 
F.
Kuban
 
P.
Electrophoresis
2017
, vol. 
38
 (pg. 
203
-
222
)
58.
Lapainis
 
T.
Scanlan
 
C.
Rubakhin
 
S. S.
Sweedler
 
J. V.
Anal. Bioanal. Chem.
2007
, vol. 
387
 (pg. 
97
-
105
)
59.
Sheeley
 
S. A.
Miao
 
H.
Ewing
 
M. A.
Rubakhin
 
S. S.
Sweedler
 
J. V.
Analyst
2005
, vol. 
130
 (pg. 
1198
-
1203
)
60.
Dailey
 
C. A.
Garnier
 
N.
Rubakhin
 
S. S.
Sweedler
 
J. V.
Anal. Bioanal. Chem.
2013
, vol. 
405
 (pg. 
2451
-
2459
)
61.
Harstad
 
R. K.
Bowser
 
M. T.
Anal. Chem.
2016
, vol. 
88
 (pg. 
8115
-
8122
)
62.
Huhn
 
C.
Ramautar
 
R.
Wuhrer
 
M.
Somsen
 
G. W.
Anal. Bioanal. Chem.
2010
, vol. 
396
 (pg. 
297
-
314
)
63.
Ramautar
 
R.
Demirci
 
A.
de Jong
 
G. J.
TrAC, Trends Anal. Chem.
2006
, vol. 
25
 (pg. 
455
-
466
)
64.
Ramautar
 
R.
Somsen
 
G. W.
de Jong
 
G. J.
Electrophoresis
2009
, vol. 
30
 (pg. 
276
-
291
)
65.
Ramautar
 
R.
Mayboroda
 
O. A.
Somsen
 
G. W.
de Jong
 
G. J.
Electrophoresis
2011
, vol. 
32
 (pg. 
52
-
65
)
66.
Monton
 
M. R.
Soga
 
T.
J. Chromatogr. A
2007
, vol. 
1168
 (pg. 
237
-
246
; discussion 236
67.
Garcia-Perez
 
I.
Vallejo
 
M.
Garcia
 
A.
Legido-Quigley
 
C.
Barbas
 
C.
J. Chromatogr. A
2008
, vol. 
1204
 (pg. 
130
-
139
)
68.
Barbas
 
C.
Moraes
 
E. P.
Villasenor
 
A.
J. Pharm. Biomed. Anal.
2011
, vol. 
55
 (pg. 
823
-
831
)
69.
R.
Ramautar
,
G. W.
Somsen
and
G. J.
de Jong
, in
Metabolomics in Practice
,
Wiley-VCH Verlag GmbH & Co. KGaA
,
2013
, pp. 177–208
70.
Ramautar
 
R.
Somsen
 
G. W.
de Jong
 
G. J.
Electrophoresis
2013
, vol. 
34
 (pg. 
86
-
98
)
71.
Elgstoen
 
K. B.
Zhao
 
J. Y.
Anacleto
 
J. F.
Jellum
 
E.
J. Chromatogr. A
2001
, vol. 
914
 (pg. 
265
-
275
)
72.
Soga
 
T.
Heiger
 
D. N.
Anal. Chem.
2000
, vol. 
72
 (pg. 
1236
-
1241
)
73.
Soga
 
T.
Ohashi
 
Y.
Ueno
 
Y.
Naraoka
 
H.
Tomita
 
M.
Nishioka
 
T.
J. Proteome Res.
2003
, vol. 
2
 (pg. 
488
-
494
)
74.
Mayboroda
 
O. A.
Neususs
 
C.
Pelzing
 
M.
Zurek
 
G.
Derks
 
R.
Meulenbelt
 
I.
Kloppenburg
 
M.
Slagboom
 
E. P.
Deelder
 
A. M.
J. Chromatogr. A
2007
, vol. 
1159
 (pg. 
149
-
153
)
75.
Ramautar
 
R.
Somsen
 
G. W.
de Jong
 
G. J.
Electrophoresis
2015
, vol. 
36
 (pg. 
212
-
224
)
76.
Ramautar
 
R.
Somsen
 
G. W.
de Jong
 
G. J.
Electrophoresis
2017
, vol. 
38
 (pg. 
190
-
202
)
77.
Kuehnbaum
 
N. L.
Kormendi
 
A.
Britz-McKibbin
 
P.
Anal. Chem.
2013
, vol. 
85
 (pg. 
10664
-
10669
)
78.
Harada
 
S.
Hirayama
 
A.
Chan
 
Q.
Kurihara
 
A.
Fukai
 
K.
Iida
 
M.
Kato
 
S.
Sugiyama
 
D.
Kuwabara
 
K.
Takeuchi
 
A.
Akiyama
 
M.
Okamura
 
T.
Ebbels
 
T. M. D.
Elliott
 
P.
Tomita
 
M.
Sato
 
A.
Suzuki
 
C.
Sugimoto
 
M.
Soga
 
T.
Takebayashi
 
T.
PLoS One
2018
, vol. 
13
 pg. 
e0191230
 
Close Modal

or Create an Account

Close Modal
Close Modal