Enhancing the Coverage of the Urinary Metabolome by Sheathless

Dec 13, 2011 - Sheathless capillary electrophoresis-mass spectrometry (CE-MS), using a porous tip sprayer, is proposed for the first time for highly s...
0 downloads 0 Views 1MB Size
Article pubs.acs.org/ac

Enhancing the Coverage of the Urinary Metabolome by Sheathless Capillary Electrophoresis-Mass Spectrometry Rawi Ramautar,*,† Jean-Marc Busnel,†,‡ André M. Deelder,† and Oleg A. Mayboroda† †

Biomolecular Mass Spectrometry Unit, Department of Parasitology, Leiden University Medical Center, Leiden, The Netherlands Beckman Coulter, Inc., Brea, California 92822, United States



S Supporting Information *

ABSTRACT: Sheathless capillary electrophoresis-mass spectrometry (CE-MS), using a porous tip sprayer, is proposed for the first time for highly sensitive metabolic profiling of human urine. A representative metabolite mixture and human urine were used for evaluation of the sheathless CE-MS platform. For test compounds, relative standard deviations (RSDs) for migration times and peak areas were below 2% and 12%, respectively, and an injection volume of only ∼8 nL resulted in detection limits between 11 and 120 nM. Approximately 900 molecular features were detected in human urine by sheathless CE-MS whereas about 300 molecular features were found with classical sheath-liquid CE-MS. This difference can probably be attributed to an improved ionization efficiency and increased sensitivity at low flow-rate conditions. The integration of transient-isotachophoresis (t-ITP) as an in-capillary preconcentration procedure in sheathless CE-MS further resulted in subnanomolar limits of detection for compounds of the metabolite mixture, and more than 1300 molecular features were observed in urine. Compared to the classical CE-MS approaches, the integration of t-ITP combined with the use of a sheathless interface provides up to 2 orders of magnitude sensitivity improvement. Hence, sheathless CE-MS can be used for in-depth metabolic profiling of biological samples, and we anticipate that this approach will yield unique information in the field of metabolomics.

C

the nebulizer gas flow may induce a hydrodynamic laminar flow toward the capillary outlet, thereby disturbing the flat flow profile usually provided by CE.7,12 In-capillary preconcentration techniques, such as pH-mediated stacking, can be used to achieve limits of detection (LODs) in the sub-μM-range for nontargeted metabolomics studies, and in combination with chemical derivatization, nanomolar level LODs for various metabolite classes can be reached.8,13,14 In order to make CE-MS a viable platform for the profiling of as many metabolites as possible in body fluids and to expand its role in the field of metabolomics, the sensitivity of CE-MS needs to be improved to enable the detection of low-abundance (highly) polar and charged metabolites in body fluids. A logical way to achieve this goal is by coupling CE to MS via a sheathless interface. Various groups developed such a sheathless interface design for CE-MS.15−18 Moini et al. demonstrated the potential of this approach for the analysis of amino acid standards, peptides, intact proteins, and protein complexes.15,16 Janini et al. and Sanz-Nebot et al. illustrated the potential of this approach for peptide analysis.17,18 In the design developed by Moini,15 the separation capillary presents an outlet end, which

apillary electrophoresis-electrospray ionization-mass spectrometry (CE-MS) is an essential analytical technique in the field of metabolomics.1−3 Over the past years, various CEMS methods have been developed for the nontargeted profiling of polar and charged metabolites in body fluids.4,5 The coupling of CE to MS can be performed via a sheath−liquid interface or a sheathless interface.6 So far, the sheath−liquid interface has been most widely used for CE-MS in metabolomics.1−5 In this configuration, the separation capillary is inserted in a tube of larger diameter in a coaxial setting. A gas flow is applied via a third coaxial capillary in order to facilitate spray formation in the ESI source. The conductive sheath liquid to which the CE terminating voltage is applied is provided via this outer tube and merges with the CE effluent at the capillary outlet. The composition and flow rate of the sheath-liquid are very important as they may influence the ESI process. CE-MS using a sheath-liquid interface is considered a robust technique; however, an intrinsic disadvantage of the use of a sheath-liquid is that it dilutes the CE effluent, thereby compromising the achievable sensitivity.7 Moreover, the ionization efficiency for some classes of metabolites along with postcapillary dilution effects may be poor when sheath-liquid CE-MS is used, resulting in detection limits that can vary over 3 orders of magnitude, i.e., from low-micromolar to millimolar range.8−11 Furthermore, it has already been demonstrated that © 2011 American Chemical Society

Received: March 1, 2011 Accepted: December 13, 2011 Published: December 13, 2011 885

dx.doi.org/10.1021/ac202407v | Anal. Chem. 2012, 84, 885−892

Analytical Chemistry

Article

mass spectrometer from Bruker (Bruker Daltonics, Bremen, Germany) via a sheathless nanospray interface.21 Separations were performed in a 100 cm long bare fused silica capillary with an internal diameter of 30 μm and an outside diameter of 150 μm. After installation, the capillary was first flushed with 0.1 M sodium hydroxide for 15 min at 1400 mbar, followed by water for 15 min at 1400 mbar and finally by the background electrolyte (BGE) of 10% acetic acid (pH 2.2) for 15 min at 1400 mbar. The BGE used in the porous separation capillary was also used as buffer in the conductive liquid capillary. Between each run, the conductive liquid capillary was rinsed for 1 min at 1400 mbar with this BGE. When not in use, the capillary was filled with the BGE and the capillary inlet and outlet were deposited in vials containing BGE and stored at room temperature. A BGE of 10% acetic acid, pH 2.2, was used for the separation. Hydrodynamic injections were used throughout the study. A voltage of +30 kV was applied during separation yielding a current of approximately 4 μA. The electrospray voltage was −1.2 kV. As usually observed in nanoESI, a very low spray voltage (−1.2 kV) was enough to form and further stabilize the spray under sheathless CE-MS conditions. Due to the observed migration time of a neutral component (dimethylsulfoxide, DMSO), the effective flow rate could be assessed to approximately 31 nL/min. For both sheathless CE-MS and sheath-liquid CE-MS, MS transfer parameters were optimized by direct infusion of an ESI tuning mix (Agilent Technologies, Waldbronn, Germany) using a LC sprayer from Agilent. Data were acquired over the mass range of 50−450 m/z with a repetition rate of 1 Hz. For sheath-liquid CE-MS, a sheath-liquid electrospray interface from Agilent Technologies (Waldbronn, Germany) was used. Separations occurred in a 100 cm long (50 μm internal diameter) fusedsilica capillary. A BGE of 10% acetic acid (pH 2.2) was used for the separation of cationic compounds. The sheath-liquid was methanol/water (1:1, v/v) containing 0.1% acetic acid and was infused using a 2.5 mL Hamilton syringe at 4 μL/min. A voltage of +30 kV was applied during separation yielding a current of approximately 10 μA. The electrospray voltage was −4.5 kV. In sheath-liquid CE-MS using a conventional ESI source the flow-rate of the sheath-liquid was 4 μL/min, so a higher voltage was required for a stable electrospray. With both interfaces, the temperature of the ion source was set to 180 °C. The porous sheathless CE-ESI interface is currently in development at Beckman Coulter, Inc. Beckman Coulter products discussed in this Article are for laboratory use only, not for use in diagnostic procedures. Test Mixture and Urine Samples. A representative metabolite mixture (carnosine, L-carnitine, creatinine, citrulline, 4-aminohippuric acid, L-tyrosine, L-phenylalanine, L-lysine, Larginine, L-isoleucine, L-leucine, L-histidine, cystathionine, homoserine, L-aspartic acid, L-hydroxyproline, hippuric acid, dopamine, phenylalanylglycine, L-cysteine, guanosine) was prepared and used to evaluate the sheathless CE-MS platform. Stock solutions of each analyte were prepared by dissolving appropriate amounts in water. Subsequently, aliquots of stock solutions were diluted with water in a 1.5 mL glass vial in order to obtain a working solution in which each analyte was present at a 100 μM concentration. Stock solutions were kept at −20 °C until usage. Human urine samples of a healthy volunteer were collected and stored at −80 °C prior to usage. Before CEMS analysis, urine samples were mixed with BGE (1:1, v/v) and centrifuged at 10 000g for 5 min. Dilution and centrifugation of the human urine sample is essential in order

has been made porous by etching with hydrofluoric acid. The electrical contact for both CE and MS is established by letting the porous capillary outlet protrude from a stainless steel ESI needle filled with static conductive liquid allowing electrospray formation at the capillary tip. Alternatively, a CE-MS interface based on a stainless steel hollow needle with a beveled sprayer tip for low flow rate CE operations can be used.19,20 Maxwell et al. demonstrated that beveled sprayer tips can effectively increase the working flow rate of an electrospray emitter, offering a performance similar to or better than that of symmetrically tapered emitters even at low flow rates.19 For a comprehensive overview on interfaces developed for CE-MS over the past decade, we refer to ref 6. Recently, the sheathless interface approach initially proposed by Moini was further developed by Beckman Coulter into a novel prototype porous tip sprayer for sheathless CE-MS. Busnel et al. have demonstrated that with this approach subnanomolar LODs can be achieved for complex peptide mixtures (protein digests).21 Using the same approach, Haselberg et al. achieved subnanomolar LODs for intact proteins, which were 50−140 times better than what can be obtained with sheath-liquid CE-MS.22 These very favorable LODs indicate that the sheathless CE-MS approach using a porous tip nanospray design shows a strong potential for highly sensitive analysis. Although not yet demonstrated, it is reasonable to expect that such a sensitivity enhancement will allow significantly higher performances when applied to metabolic profiling of body fluids by CE-MS, thus ultimately improving the achievable coverage of metabolites. If so, this would be of pivotal importance for obtaining in-depth insight into disease biochemistry and to discover potentially novel metabolic biomarkers. Therefore, the aim of this study was to evaluate the performance of this novel sheathless CE-MS platform for highly sensitive metabolic profiling of body fluids. Using a representative metabolite mixture, we first evaluated analytical parameters such as migration time and peak area repeatability, response linearity, and LODs. The applicability of the sheathless CE-MS method for highly sensitive metabolic profiling of body fluids was then investigated for human urine. The use of transient-isotachophoresis (t-ITP) as an incapillary preconcentration technique was also assessed. Results were compared to sheath-liquid CE-MS to determine whether, and to what extent, an improved coverage of metabolites in urine could be obtained with sheathless CE-MS.



MATERIALS AND METHODS Chemicals. All chemicals used were of analytical grade or higher purity. Carnosine, L-carnitine, creatinine, citrulline, 4aminohippuric acid, L-tyrosine, L-phenylalanine, L-lysine, Larginine, L-isoleucine, L-leucine, L-histidine, cystathionine, homoserine, L-aspartic acid, L-hydroxyproline, hippuric acid, dopamine, phenylalanylglycine, guanosine, acetic acid, and ammonium acetate were purchased from Fluka (Buchs, Switzerland). CE-ESI-MS. The CE experiments were carried out with a prototype capillary electrophoresis (CE) system from Beckman Coulter (Brea, California, USA) equipped with a temperature controlled autosampler and a power supply able to deliver up to 30 kV. Bare fused-silica capillaries etched with a porous tip were made available by Beckman Coulter (Brea, California, USA); for details, we refer to ref 21. CE-ESI-MS experiments were performed on the prototype instrument coupled to a microTOF 886

dx.doi.org/10.1021/ac202407v | Anal. Chem. 2012, 84, 885−892

Analytical Chemistry

Article

wide variety of cationic metabolites, and a relatively low conductivity, we used the same BGE in the present study. The analysis of the representative metabolite test mixture by sheathless CE-MS yielded excellent analyte intensities as shown in Figure 1. Compounds that are doubly positively

to prevent clogging of the porous tip capillary by particulate matter in urine. The repeatability of the sheathless CE-MS method was determined by eight consecutive analyses of the representative metabolite mixture (1 μM) and of human urine spiked with test compounds (1 μM). Linearity of response for test compounds in 1:1 diluted BGE was evaluated by measurement of six (or eight) analyte concentrations ranging from 0.1 (or 0.5) to 25 μM. Linearity of response for test compounds spiked into human urine was evaluated by measurement of five analyte concentrations ranging from 0.1 (or 0.5) to 25 μM in triplicate. Peak areas for spiked compounds in urine were corrected for endogenous levels of the test compounds. For the determination of the batch-tobatch repeatability, three individual porous tip capillary sprayers were used. Batch-to-batch precision was evaluated by determining the relative standard deviations (RSDs) of migration times and peak areas of five test compounds spiked into human urine at a concentration of 1 μM using eight repeated injections on each capillary (peak areas for spiked compounds in urine were corrected for endogenous levels of the test compounds). For sheath-liquid CE-MS, linearity of response was determined by measurement of five analyte concentrations ranging from 1 (or 5) to 150 (or 200) μM. All urine samples were diluted with BGE (1:1, v/v) prior to analysis by CE-MS. Data Analysis. CE-MS data were analyzed using Bruker Daltonics Data Analysis software. Peak areas of test compounds were determined from extracted ion electropherograms. The number of molecular features (i.e., the number of peaks detected above a certain intensity threshold within the CE run time) in a sample was determined using the following parameters in the “Find Molecular Features” function in the DataAnalysis software: S/N threshold = 5; correlation coefficient threshold = 0.95; minimum compound length = 5; smoothing width = 2. After this procedure, interfering components such as polymer molecules, probably originating from solvent vials, and impurities present in the solutions used as well as fragments and adducts were manually excluded. In order to avoid accounting for the salts clusters observed in sheathless and sheath-liquid CE-MS, the number of molecular features in human urine was determined for the migration time range 8−30 min. For the provisional identification of metabolites, rational chemical formulas were generated on the basis of internally calibrated mass spectra using the SmartFormula tool within the Data Analysis software. The chemically reasonable formulas were submitted to metabolome databases: the Human Metabolome Database (HMDB)23 and the METLIN database.24 Isotopic distribution patterns of the matched metabolite candidates were then simulated with the Simulate Pattern tool (Data Analysis software) and compared with observed mass spectra to further reduce the number of potential elemental compositions.

Figure 1. Multiple extracted ion electropherogram for the metabolite test mixture (5 μM) obtained with sheathless CE-MS using a porous tip sprayer. Peaks: 1, carnosine; 2, creatinine; 3, lysine; 4, arginine; 5, histidine; 6, carnitine; 7, dopamine; 8, phenylalanylglycine; 9, isoleucine; 10, leucine; 11, cystathionine; 12, homoserine; 13, citrulline; 14, phenylalanine; 15, tyrosine; 16, cystine; 17, aspartic acid; 18, 4-aminohippuric acid; 19, hydroxyproline; 20, guanosine; 21, hippuric acid. Conditions: BGE, 10% acetic acid (pH 2.2); sample injection, 2.0 psi for 30 s.

charged at pH 2.2, such as carnosine, lysine, arginine, and histidine, migrated first, followed by singly positively charged compounds, such as leucine, tyrosine, and guanosine (Figure 1). Hippuric acid, which is slightly negatively charged at pH 2.2, migrated directly after the migration time of the neutral marker (ca. 23 min for DMSO, see Materials and Methods). Under these conditions, isoleucine could be partly separated from leucine and plate numbers ranged from 50 000 to 200 000 for the test compounds, indicating acceptable separation efficiency. Next, the repeatability of the method was examined by eight consecutive injections of the test mixture (1 μM of each compound). Favorable RSDs for migration times (