Fluorescent derivatization combined with aqueous solvent-based dispersive liquid-liquid microextraction for determination of butyrobetaine, l -carnitine and acetyl-l -carnitine in human plasma

G Model

ARTICLE IN PRESS

CHROMA-357830; No. of Pages 10

Journal of Chromatography A, xxx (2016) xxx–xxx

Contents lists available at ScienceDirect

Journal of Chromatography A journal homepage: www.elsevier.com/locate/chroma

Fluorescent derivatization combined with aqueous solvent-based dispersive liquid-liquid microextraction for determination of butyrobetaine, l-carnitine and acetyl-l-carnitine in human plasma Yi-Ching Chen a , Chia-Ju Tsai a , Chia-Hsien Feng a,b,c,∗ a

Department of Fragrance and Cosmetic Science, College of Pharmacy, Kaohsiung Medical University, Kaohsiung 80708, Taiwan Ph.D. Program in Toxicology, College of Pharmacy, Kaohsiung Medical University, Kaohsiung 80708, Taiwan c Institute of Medical Science and Technology, National Sun Yat-sen University, Kaohsiung 80424, Taiwan b

a r t i c l e

i n f o

Article history: Received 21 April 2016 Received in revised form 10 August 2016 Accepted 11 August 2016 Available online xxx Keywords: Aqueous solvent-based dispersive liquid-liquid microextraction l-carnitine Fluorescence derivatization Narrow-bore liquid chromatography

a b s t r a c t A novel aqueous solvent-based dispersive liquid-liquid microextraction (AS-DLLME) method was combined with narrow-bore liquid chromatography and fluorescence detection for the determination of hydrophilic compounds. A remover (non-polar solvent) and extractant (aqueous solution) were introduced into the derivatization system (acetonitrile) to obtain a water-in-oil emulsion state that increased the mass transfer of analytes. As a proof of concept, three quaternary ammonium substances, including butyrobetaine, l-carnitine and acetyl-l-carnitine, were also used as analytes and determined in pharmaceuticals, personal care products, food and human plasma. The analytes were derivatized with 4-bromomethylbiphenyl for fluorescence detection and improved retention in the column. The linear response was 10–2000 nM for l-carnitine and acetyl-l-carnitine with a good determination coefficient (r2 > 0.998) in the standard solution. The detection limit for l-carnitine and acetyl-l-carnitine was 4.5 fmol. The method was also successfully applied to a 1 ␮L sample of human plasma. In the linearity calculations for determining butyrobetaine, l-carnitine and acetyl-l-carnitine in human plasma, the determination coefficients ranged from 0.996 to 0.999. Linear regression exhibited good reproducibility and a relative standard deviation better than 7.50% for the slope and 9.06% for the intercept. To characterize highly hydrophilic compounds in various samples, the proposed method provides good sensitivity for a small sample volume with a low consumption of toxic solvents. © 2016 Elsevier B.V. All rights reserved.

1. Introduction Dispersive liquid-liquid microextraction (DLLME) was first introduced by Rezaee et al. to provide a simple and miniaturized technique for sample pretreatment [1]. Various modifications of DLLME have been proposed to improve its versatility and applicability. For example, DLLME uses a dispersant to induce the formation of fine droplets for improved liquid-liquid extraction efficiency. The extractant that was used in early applications of DLLME was a small amount of chlorinated solvent with a density higher than that of water. Although the use of high-density solvents increases the convenience of DLLME by enhancing sedimentation, high-density solvents have harmful environmental and human health effects. The proposed solution was low-density DLLME

∗ Corresponding author at: 100, Shih Chuan 1st Road, Kaohsiung 80708, Taiwan. E-mail addresses: [email protected], [email protected] (C.-H. Feng).

(LD-DLLME), in which a low-density organic solvent is used rather than chlorinated solvents. However, one disadvantage is that the extraction phase forms on top of the aqueous phase in LD-DLLME. To withdrawal the organic phase, various apparatuses have been developed to enable convenient handling of small supernatant volumes [2–7]. The DLLME and LD-DLLME methods are primarily used to extract hydrophobic compounds from water samples and difficult to use for separating polar analytes from aqueous samples. Therefore, an effective method for extracting extremely hydrophilic analytes, such as l-carnitine (2.50 g mL−1 in water at 20 ◦ C) [8], from aqueous samples is needed. In this study, we developed a novel approach for using an aqueous solvent (AS) in DLLME. After using a small volume of water (or AS) as the extractant to obtain the hydrophilic analytes, a derivatization medium (acetonitrile, ACN) was used as a dispersant to assist the formation of a water-in-oil emulsion state. The use of a low-density organic solvent as the continuous phase and remover also improved the phase separation and removed hydrophobic

http://dx.doi.org/10.1016/j.chroma.2016.08.030 0021-9673/© 2016 Elsevier B.V. All rights reserved.

Please cite this article in press as: Y.-C. Chen, et al., Fluorescent derivatization combined with aqueous solvent-based dispersive liquidliquid microextraction for determination of butyrobetaine, l-carnitine and acetyl-l-carnitine in human plasma, J. Chromatogr. A (2016), http://dx.doi.org/10.1016/j.chroma.2016.08.030

G Model CHROMA-357830; No. of Pages 10

ARTICLE IN PRESS Y.-C. Chen et al. / J. Chromatogr. A xxx (2016) xxx–xxx

2

interferences. Then, the two phases were separated by centrifugation, and the bottom of the aqueous phase was collected for direct analysis. The proposed AS-DLLME method is applicable for analyzing hydrophilic analytes that are difficult to extract from aqueous samples. The method does not require the use of any additional apparatus or the use of hypertoxic organic solvents. As a proof of concept, three quaternary ammonium substances were selected as analytes including butyrobetaine, l-carnitine and acetyl-l-carnitine. l-carnitine is essential for fatty acid oxidation in the human body [9]. Butyrobetaine is a key precursor of carnitine biosynthesis in the kidney, liver and brain [10]. Acetyl-l-carnitine is an l-carnitine derivative that delays aging-associated degeneration of brain function, cognition and memory [11]. The ratio of acyl-l-carnitine to l-carnitine is considered a useful indicator of the state of health, and l-carnitine depletion has been associated with various diseases. Therefore, an effective method for monitoring the concentrations of l-carnitine and its related compounds is needed. The previously reported l-carnitine determination methods include colorimetry and radioenzymatic assay [12–14], capillary electrophoresis (CE) [15–17], gas chromatography (GC) [18] and high performance liquid chromatography (HPLC) [19–28]. These methods typically require derivatization of butyrobetaine, l-carnitine and acetyl-l-carnitine for several reasons. First, because these compounds are hydrophilic, derivatization may reduce the polarity and improve retention factors in GC and reverse-phase HPLC. Second, providing chromophores or fluorophores for these three analytes can enhance the detection sensitivity during ultraviolet (UV) detection or fluorescence detection (FLD). However, one of the most common methods of clinical analysis of l-carnitine and its related compounds is HPLC combined with tandem mass spectrometry (MS/MS). In this method, butanolic HCl is added to biosamples, such as dried blood (15.2 ␮L) extractant, plasma, serum, or urine (25–100 ␮L) [29–36]. After derivatization, the solution is dried and dissolved in a mobile phase for HPLC combined with MS/MS, which is more expensive and more difficult to maintain compared to UV and FLD. The objectives of this study were as follows: (1) to develop a rapid, sensitive and reliable analytical process that combines derivatization and narrow-bore LC with FLD for increased analytical sensitivity, (2) to use butyrobetaine, l-carnitine and acetyl-l-carnitine to demonstrate that AS-DLLME increases the extractability of water-soluble analytes in water samples and (3) to validate the application of the proposed method for use with pharmaceuticals, personal care products, food and human plasma.

2. Materials and methods

100 ␮M and water at concentrations of 200 ␮M for standard and human plasma analysis, respectively. The different concentrations of the working solution were obtained by diluting the stock solution. The Br-MBP solution was freshly prepared in ACN. KOH, K2 CO3 , and KHCO3 were prepared in ACN, and IS was prepared in MeOH. All of the solutions were stored at 4 ◦ C prior to use. l-carnitine chewable tablets and injections (CARNITENE® ) that were purchased from Harvest & Health CO., LTD. (Taiwan) were analyzed. Personal care products and food samples containing l-carnitine were purchased from local markets and included capsules, candy, nutritional drinks, coffee, gels, shampoos, and creams. The samples were maintained in their original containers prior to analysis. The plasma samples from three healthy blood donors were collected in heparin tubes. The experiments were approved by the Institutional Review Board of Kaohsiung Medical University Chung-Ho Memorial Hospital (KMUH-IRB-20130135). 2.2. Instruments The narrow-bore LC-FLD system consisted of an Agilent 1200 LC system (Santa Clara, CA, USA) with binary pump, degasser, autosampler and 1260 fluorescence detector. The analytical column was a Chromolith® PerformanceRP-18e column (100 mm × 2 mm), which was obtained from Merck. The nano LC–MS/MS system consisted of a Waters nano ACQUITYUPLC system (Milford, MA, USA) and LTQ Orbitrap Discovery hybrid Fourier Transform Mass Spectrometer (Thermo Fisher Scientific, Inc. Bremen, Germany) with a nano spray source and a resolution of 30000. The desalting column (Symmetry C18; 180 ␮m × 20 mm, 5 ␮m) and analytical column (BEH C18; 75 ␮m × 100 mm, 1.7 ␮m) were obtained from Waters (Milford, MA, USA). 2.3. Sample preparation for pharmaceuticals, personal care products, and food items The analysis included fourteen samples (i.e., pharmaceuticals, personal care products and food items) of varying formulations. Depending on the complexity of their content, different sample preparations were used. For tablets, 10 units were ground and homogenized. Then, a 0.5 mg sample of the powder was diluted in ACN up to the calibration range. For personal care products and food samples, 10 mg samples were diluted in ACN up to the calibration range. For injection formulas, 10 units were mixed, and 10 ␮L of the injection formula were diluted in ACN up to the calibration range. All of the ACN solutions were filtered through 0.45 ␮m filters. Finally, 20 ␮L of the solution was derivatized with Br-MBP for all of the samples.

2.1. Chemicals, reagents and materials 2.4. Sample preparation for human plasma Acetonitrile, methanol (MeOH), dimethyl sulfoxide (DMSO), acetone, toluene, n-hexane, ethyl acetate (EA), ammonium acetate (NH4 OAC), ammonium chloride (NH4 Cl), sodium sulfate (Na2 SO4 ), sodium chloride (NaCl), ammonium bicarbonate (NH4 HCO3 ), sodium bicarbonate (NaHCO3 ), potassium hydroxide (KOH), potassium carbonate (K2 CO3 ), potassium bicarbonate (KHCO3 ), 25% ammonia solution (NH4 OH), formic acid (FA), and acetic acid (AA) were purchased from Merck (Darmstadt, Germany). l-carnitine, 4-bromomethylbiphenyl (Br-MBP), and 9-aminoacridine (internal standard, IS) were purchased from the Tokyo Chemical Industry (Tokyo, Japan). Butyrobetaine hydrochloride and acetyl-l-carnitine hydrochloride were obtained from Sigma-Aldrich (St. Louis, MO, USA). Distilled water was obtained from a Millipore Milli-Q system (Bedford, USA). The stock solutions of butyrobetaine, l-carnitine, and acetyl-lcarnitine were prepared by dissolution in ACN at concentrations of

A 1 ␮L aliquot of human plasma was pipetted into PCR tubes containing 20 ␮L of analytes in varying amounts of the aqueous solution (0, 5, 25, 100, 200 pmol). The protein precipitating reagent (85 ␮L volume of acetone) was vortexed with the sample solution for 2 min and centrifuged for 10 min at 14800 rpm. Then, 100 ␮L of the supernatant was removed and dried with a centrifugal evaporator at 37 ◦ C. The residue was used in subsequent derivatization steps. 2.5. Derivatization methods The derivatization steps for the standard and pharmaceutical, personal care product, and food product samples were as follows: First, 20 ␮L of the ACN solution described in Section 2.3 was transferred to PCR tubes. Next, 5 ␮L of the Br-MBP solution (15 mM in

Please cite this article in press as: Y.-C. Chen, et al., Fluorescent derivatization combined with aqueous solvent-based dispersive liquidliquid microextraction for determination of butyrobetaine, l-carnitine and acetyl-l-carnitine in human plasma, J. Chromatogr. A (2016), http://dx.doi.org/10.1016/j.chroma.2016.08.030

G Model CHROMA-357830; No. of Pages 10

ARTICLE IN PRESS Y.-C. Chen et al. / J. Chromatogr. A xxx (2016) xxx–xxx

3

Table 1 Optimal derivatization conditions for standards and human plasma. Parameter

Concentration of Br-MBP (mM) Volume of Br-MBP (␮L) Reaction temperature (◦ C) Reaction time (min)

Standards

Human plasma

Studied ranges

Optimal condition

Studied ranges

Optimal condition

2–35 3–13 80–100 3–13

15 5 90 7

0.5–9 15–65 80–100 3–13

5 25 90 9

ACN) and 2 ␮L of a KOH solution (saturated in ACN) were added. After vortexing, the solution was heated in a dry bath incubator at 90 ◦ C for 7 min. For human plasma, the residue was dissolved in 25 ␮L of the Br-MBP solution (5 mM in ACN). After thorough vortexing for 1 min, the solution was allowed to react for 9 min at 90 ◦ C. After derivatization, all of the reaction solutions were cooled to room temperature prior to performing the optimized AS-DLLME procedure. Table 1 shows the optimized parameters and factors that affected the derivatization efficiency. 2.6. AS-DLLME After the derivatization, 85 ␮L of toluene and 5 ␮L of H2 O (for all samples except human plasma) or 1 M NH4 OAC (for human plasma) were added to the solution and vortexed to obtain a water-in-oil emulsion. The sediment phase (5 ␮L) was obtained by centrifuging the water-in-oil emulsion for 3 min at 14800 rpm. Finally, 5 ␮L of IS was added to the collected sediment phase and injected into a narrow-bore LC-FLD for quantitation as well as a nano LC–MS/MS system for identification. 2.7. Narrow-bore LC-FLD condition The mobile phases consisted of eluent A (0.2% formic acid in water) and eluent B (MeOH) at a flow rate of 0.4 mL min−1 . The standard, pharmaceuticals, personal care products, and food products were analyzed under the following gradient conditions: 0–0.3 min, 25–35% B; 0.3–1 min, 35–50% B; 1–6.5 min, 50–60% B. The injection volume for these sample solutions was 1.5 ␮L. The excitation wavelength was set to 255 nm. The emission wavelength was set to 500 nm from 0 to 4.8 min and then set to 317 nm until the end of the analysis for detection of IS and analytes, respectively. For the plasma sample, the gradient conditions were as follows: 0–1 min, 25–35% B; 1–10 min, 35–45% B; 10–12 min, 45–50% B; 12–12.5 min, 50% B. The injection volume was set to 3 ␮L. The excitation and emission wavelengths as described above except the wavelength was switched at 5.01 min. 2.8. Nano LC–MS/MS condition The mobile phases of solvents A and B consisted of 0.1% FA and ACN containing 0.1% FA, respectively, at a flow rate of 300 nL min−1 . After desalting for 3 min with 0.1% FA at 5 ␮L min−1 , the separation gradient conditions were as follows: 0–1 min, 1–12% B; 1–5 min, 12–100% B; 5–50 min, 100% B. The source was operated in positive ionization mode. The mass scan range was 100–1000 m/z. (aq)

3. Results and discussion 3.1. Optimization of derivatization To increase the retention and detection sensitivity, derivatization was performed with Br-MBP. All of the derivatization parameters were optimized including the medium, concentration and volume of Br-MBP, the type of base and the temperature and

duration of the reaction. Table 1 lists the optimal parameters. All of the experimental results were performed in triplicate, and average values are reported.

3.1.1. Optimization of the derivatization of standards The carboxyl group of l-carnitine and acetyl-l-carnitine preceded the nucleophilic substitution reaction with Br-MBP. Polar protic solvents containing the OH or NH group decelerated the SN 2 reactions due to solvation of the reactant nucleophile. Aprotic solvents are preferred because they reduce the free energy, which accelerates the SN 2 reactions. Three aprotic solvents (i.e., ACN, acetone and DMSO) were evaluated for use in the derivatization procedure. In the derivatization medium, acetone and DMSO had the lowest responses. Acetone exhibited a low response due to its aldol reaction during the dissolution in KOH [37]. DMSO exhibited a low response due to the poor solubility of Br-MBP (less than 25 mM). Therefore, ACN was used in subsequent experiments. Then, the solutions with different Br-MBP concentrations and volumes were compared. Derivatization efficiency incrementally increased at Br-MBP concentrations of 2–15 mM and plateaued at Br-MBP concentrations of 15–35 mM (Fig. S1 in Supplemental Data). Because the Br-MBP in the ACN solution was also used as a dispersant in the AS-DLLME system, the Br-MBP volume influenced the extraction efficiency. Fig. S2 (Supplemental Data) shows the results of comparisons of Br-MBP (15 mM) with various volumes (3–13 ␮L). The comparisons indicated that 3 ␮L of Br-MBP provided an inadequate derivatization efficiency but 5 ␮L of Br-MBP exhibited the highest efficiency. However, when the Br-MBP volume exceeded 5 ␮L, the response decreased for two reasons. First, an increase in the Br-MBP volume decreased the final analyte concentration, which reduced the reactivity. Second, the volume of the dispersant (ACN) increased when the Br-MBP volume increased. Excessive dispersant in the AS-DLLME reduced the extraction efficiency because it enhanced the solubility of analytes in the remover. Therefore, 5 ␮L of 15 mM Br-MBP had the highest derivatization and extraction efficiencies, and this dispersant quantity was used in subsequent experiments. Next, KOH, K2 CO3 and KHCO3 were evaluated as bases for use in the derivatization. In order from highest to lowest, the derivatization efficiencies were KOH, K2 CO3 , KHCO3 , and no base (Fig. S3 in Supplemental Data). The KOH solution provided an appropriate alkaline environment for derivatization. This solution maintained the carboxylic group in an ionic state because the derivatives were obtained by the nucleophilic reagents attacking the carbon atom of Br-MBP. Therefore, KOH was considered the optimal base for derivatization. The results from varying the reaction temperatures (80, 90 and 100 ◦ C) and derivatization times (range from 3 to 13 min) were also compared. The maximum yield was achieved at 90 ◦ C and 7 min (Fig. S4 in Supplemental Data). Lower reaction temperatures required a longer period of time to achieve the same response. Higher reaction temperatures required nearly the same time to obtain a similar response. Except for human plasma, derivatization of all of the samples was performed at 90 ◦ C for 7 min.

Please cite this article in press as: Y.-C. Chen, et al., Fluorescent derivatization combined with aqueous solvent-based dispersive liquidliquid microextraction for determination of butyrobetaine, l-carnitine and acetyl-l-carnitine in human plasma, J. Chromatogr. A (2016), http://dx.doi.org/10.1016/j.chroma.2016.08.030

G Model CHROMA-357830; No. of Pages 10 4

ARTICLE IN PRESS Y.-C. Chen et al. / J. Chromatogr. A xxx (2016) xxx–xxx

Fig. 1. Factors that affect efficiency in extracting l-carnitine, butyrobetaine and acetyl-l-carnitine derivatives from standard solution (line) and from human plasma (bar): (a) toluene volume, (b) extractant type, (c) H2 O volume, (d) NH4 OAc concentration, (e) NH4 OAc volume. Error bars represent the standard deviation. C: l-carnitine; B: butyrobetaine; AC: acetyl-l-carnitine.

3.1.2. Optimization of derivatization for human plasma All of the derivatization parameters that were investigated for the standards were re-examined to determine whether the complexity of the biosample matrix would alter the optimal derivatization conditions. To maintain the same ACN volume (dispersant) in AS-DLLME, 25 ␮L of Br-MBP were used, and Br-MBP concentrations of 0.5–9 mM were tested. The derivatization efficiency increased as the Br-MBP concentration increased from 0.5 mM to 5 mM. At Br-MBP concentrations higher than 5 mM, the derivatization efficiency exhibited no additional increase (Fig. S5 in Supplemental Data). Therefore, 5 mM Br-MBP was considered optimal. The residue described in Section 2.4 was used to compare the derivatization efficiency among ACN, ACE and DMSO. For human plasma, ACN still resulted in the best results. Therefore, ACN was used as the derivatization medium in all of the experiments. Alkaline conditions promoted the reaction. The investigated bases included KOH, K2 CO3 and KHCO3 saturated in ACN. The results in Fig. S6 (Supplemental Data) indicate that no effect was produced by the addition of alkaline agents, which is most likely due to the plasma sample itself being a weak alkaline environment that was sufficient for derivatization. Therefore, no basic agent was included in the optimum conditions. The results for reaction temperatures of 80 ◦ C–100 ◦ C and reaction times of 3–13 min were compared. Heating at 90 ◦ C for 9 min resulted in the highest derivatization efficiency (Fig. S7 in Supplemental Data). A lower energy was inadequate, and the signal slowly degraded when the temperature exceeded 90 ◦ C and the reaction time exceeded 9 min. In comparison to the standards, human plasma had a longer reaction time because it contained many carboxylic compounds, such as amino acids and fatty acids. Therefore, the formation of derivatives consumes more energy.

3.2. Optimization of AS-DLLME In AS-DLLME, a remover (non-polar solvent) and extractant (aqueous solution) were introduced into the derivatization system (ACN) to prepare a water-in-oil emulsion that can accelerate mass transfer during extraction. Then, demulsification was performed by centrifugation, and the sediment consisted of the analyte-enriched phase. The supernatant was the remover that could remove the hydrophobic interferences (e.g., excess Br-MBP). The sediment at the bottom can be directly analyzed. To maximize the extractability of l-carnitine and its relative compounds in water samples, various AS-DLLME parameters were experimentally compared and optimized including the type and volume of remover as well as the type, volume, and concentration of the extractant. The peak area ratio of each analyte and IS was reported as the average of three replicates with error bars that represent the standard deviation.

3.2.1. Optimization of AS-DLLME for a standard solution For a convenient and environmentally friendly procedure, a small amount of low-density solvent was used as the remover in AS-DLLME to obtain a water-in-oil emulsion state and prevent interference from the sample matrix. The tests employed EA, toluene and hexane. EA did not generate a sediment phase, and hexane produced a three-layer interface. Only toluene resulted in an emulsion, and a sediment phase was easily obtained after centrifugation. Therefore, toluene was selected as the remover. An appropriate amount of the remover in a well-dispersed system can produce efficient mass transfer in AS-DLLME. If the amount of remover is insufficient, the dispersant makes the extractant and remover miscible, and an emulsion cannot be achieved. However, excessive remover reduces the clouding phenomenon by reducing the contact area between the analytes and the extractant, which subsequently reduces mass transfer efficiency and extraction efficiency. The efficiency for the extraction of the derivatives was

Please cite this article in press as: Y.-C. Chen, et al., Fluorescent derivatization combined with aqueous solvent-based dispersive liquidliquid microextraction for determination of butyrobetaine, l-carnitine and acetyl-l-carnitine in human plasma, J. Chromatogr. A (2016), http://dx.doi.org/10.1016/j.chroma.2016.08.030

G Model CHROMA-357830; No. of Pages 10

ARTICLE IN PRESS Y.-C. Chen et al. / J. Chromatogr. A xxx (2016) xxx–xxx

5

Fig. 2. Chromatograms for determining l-carnitine and acetyl-l-carnitine: (a) blank (dashed line) and 2-␮M standard solution (solid line), (b) pharmaceutical A, (c) food sample C, (d) personal care product A. Peak 1: l-carnitine derivative; peak 2: acetyll-carnitine derivative; I.S.: 9-aminoacridine.

compared for different volumes (55–105 ␮L) of the remover (i.e., toluene). The results in Fig. 1(a) indicate that an emulsion was formed throughout this range. The extraction efficiency increased as the toluene volume increased from 55 to 85 ␮L, and 85 ␮L of toluene resulted in the maximum extraction efficiency. Toluene volumes that exceeded 85 ␮L resulted in no further changes in the extraction efficiency. Therefore, 85 ␮L of toluene was considered optimal. The extractant is another crucial factor for the extraction efficiency. Because the structures of l-carnitine and its related compounds contain quaternary ammonium, which is highly soluble in water, the extraction of these analytes from aqueous samples is challenging. The AS-DLLME in this study used aqueous reagents as extractants to address the low extractability limitation of organic solvents. Seven aqueous reagents including water, Na2 SO4(aq) , NH4 Cl (aq) , NH4 OAc (aq) , NaCl (aq) , NH4 HCO3(aq) , and NaHCO3(aq) were compared. Na2 SO4(aq) and NaHCO3(aq) resulted in precipitation during the extraction process. Fig. 1(b) shows the experimental results for the five remaining reagents. The extractant that produced the best peak area ratio was water followed by NH4 OAc (aq) , NH4 Cl (aq) , NaCl (aq) and NH4 HCO3(aq) . Therefore, the extraction efficiency was compared for varying volumes (5–15 ␮L) of water that was used as the extractant. 5–15 ␮L of water resulted in the formation of a water-in-oil emulsified state. However, additional increases in the volume of water did not increase the extraction efficiency because the increased dilution of the analytes lowered their responses (Fig. 1(c)). Low extractant volumes resulted in the most efficient enrichment. Therefore, 5 ␮L of water resulted in the optimal extraction efficiency. Fig. 2(a) shows the chromatogram for 2 ␮M standard solutions after derivatization and AS-DLLME. 3.2.2. Optimization of AS-DLLME for human plasma A human plasma biosample has a complex composition. The main factors for using AS-DLLME for human plasma were optimized to minimize interference. In tests using EA, toluene and hexane, the extraction results for human plasma were similar to those for the standards. Toluene was the optimal remover. Fig. 1(a) compares the extraction efficiency at volumes ranging from 55 to 105 ␮L. At toluene volumes of 55–85 ␮L, the extraction efficiency increased with the volume. When the toluene volume exceeded 85 ␮L, the extraction efficiency did not further increase. Therefore, 85 ␮L of toluene was still considered the optimal volume for human plasma. Next, five aqueous solvents (i.e., water, NH4 Cl (aq) , NH4 OAc (aq) , NaCl (aq) and NH4 HCO3(aq) ) were experimentally compared. In contrast to the standards, the results in Fig. 1(b) indicate that the best extractant in this experiment was NH4 OAc (aq) rather than

Fig. 3. Chromatograms for determining butyrobetaine, l-carnitine and acetyl-lcarnitine derivatives in human plasma: (a) dashed line indicates plasma spiked with 200 pmol standard solution and extracted by AS-DLLME; solid line indicates plasma spiked with 200 pmol standard solution without extraction, (b) plasma blank (dashed line) and plasma spiked with 200 pmol standard solution (solid line). Peak 1: l-carnitine derivative; peak 2: butyrobetaine derivative; peak 3: acetyl-l-carnitine derivative; peak I.S.: 9-aminoacridine.

water because the NH4 OAc (aq) solution had a pH of 7 in human plasma matrix. These observations resulted from differences in the standards. Therefore, NH4 OAc (aq) was used as the extractant in the AS-DLLME for human plasma. Fig. 1(d) compares the results for NH4 OAc concentrations ranging from 0.1 to 1000 mM. Because the signal increased with the NH4 OAc concentration, 1000 mM was selected as the NH4 OAc (aq) extractant concentration for use in subsequent experiments. The effects of the extractant volume were investigated by comparing 1000 mM NH4 OAc with varying volumes (5–13 ␮L). The results in Fig. 1(e) indicate that the peak area ratio increased as the NH4 OAc volume decreased because a small extractant volume produced a high enrichment factor. However, when the volume of the NH4 OAc solution was less than 5 ␮L, an insufficient sediment phase was obtained. Therefore, 5 ␮L of 1000 mM NH4 OAc was chosen as the optimal condition. Fig. 3(a) shows the chromatograms for human plasma spiked with the 200 pmol standard solution. The dashed line indicates human plasma extracted by AS-DLLME, and the solid line indicates human plasma not subjected to the extraction procedure. The two chromatograms show the two advantages of AS-DLLME. First, AS-DLLME exhibits a better analyte response, which increased sensitivity. Second, the sample was not treated by AS-DLLME. The excess derivatizing agent was eluted from the column after 15 min of chromatographic analyses. Therefore, removal of the excess derivatization reagent in the extraction process reduced the analysis time to 10 min, reduced consumption of the mobile phase, and prevented column contamination and plugged packing. In summary, the AS-DLLME method is superior in terms of the detection limit, column life, and cost effectiveness.

Please cite this article in press as: Y.-C. Chen, et al., Fluorescent derivatization combined with aqueous solvent-based dispersive liquidliquid microextraction for determination of butyrobetaine, l-carnitine and acetyl-l-carnitine in human plasma, J. Chromatogr. A (2016), http://dx.doi.org/10.1016/j.chroma.2016.08.030

G Model

ARTICLE IN PRESS

CHROMA-357830; No. of Pages 10

Y.-C. Chen et al. / J. Chromatogr. A xxx (2016) xxx–xxx

6

Table 2 Analytical performance of the proposed method for l-carnitine and acetyl-l-carnitine. Analytes

Linearity (nM)

Linear equation (n = 6)

Precision (RSD,%)

Accuracy (RE,%)

30 nM

500 nM

1500 nM

30 nM

500 nM

1500 nM

Intraday l-carnitine acetyl-l-carnitine

10–2000 10–2000

y = 0.7464 × −0.0161 y = 0.9516 × −0.0386

3.43 3.97

3.60 4.28

4.75 3.03

4.71 3.54

1.85 0.89

3.54 4.16

Interday l-carnitine acetyl-l-carnitine

10–2000 10–2000

y = 0.7394 × −0.0173 y = 0.8973 × −0.0044

3.80 4.91

3.50 4.11

3.94 4.97

4.60 5.08

−0.86 −5.47

2.66 1.77

Table 3 Relative recoveries of l-carnitine pharmaceuticals. Amount spiked (␮M)

Amount found (␮M)

RSD (%)

Recovery (%)

Pharmaceutical A (tablet) 0 0.05 0.5 1

4.7 × 10-1 5.2 × 10-1 9.8 × 10-1 1.5 × 100

4.30 1.43 1.97 0.80

– 100.0 102.0 104.0

Pharmaceutical B (injection) 0 5.9 × 10-1 6.4 × 10-1 0.05 1.1 × 100 0.5 1.6 × 100 1

2.22 2.09 1.37 2.99

– 100.0 100.0 104.0

Table 4 Linear regression equations for intraday and interday assays in human plasma. Analytes

Slope (RSD,%)

Intercept (RSD,%)

r2

Intraday (n = 6) butyrobetaine l-carnitine acetyl-l-carnitine

0.0058 (5.17) 0.0075 (6.67) 0.0040 (5.00)

0.0145 (6.90) 0.3362 (5.53) 0.0254 (7.09)

0.997 0.999 0.999

Interday (n = 6) butyrobetaine l-carnitine acetyl-l-carnitine

0.0054 (7.41) 0.0069 (5.80) 0.0040 (7.50)

0.0069 (8.70) 0.3132 (6.55) 0.0254 (9.06)

0.996 0.999 0.999

3.3. Evaluation of the method The method was validated by generating calibration curves for lcarnitine and acetyl-l-carnitine standards in a concentration range from 10 to 2000 nM. The peak area ratios of the analytes to IS were calculated and evaluated by linear regression (Table 2). In the interday analyses, the linear regression determination coefficients for l-carnitine and acetyl-l-carnitine were 0.999 and 0.998, respectively. The limit of detection (LOD) was measured to be 3 times the signal-to-noise ratio, which corresponds to 4.5 fmol for l-carnitine and acetyl-l-carnitine. The precision and accuracy were calculated for the intraday and interday assays (n = 6) that were performed at three different concentrations (i.e., 30, 500, and 1500 nM), which represented the concentration range of the calibration standards. Table 2 shows the good precision and accuracy that was obtained using the derivatization in combination with the AS-DLLME method. The recoveries of l-carnitine from pharmaceuticals A and B were evaluated with samples spiked with three different levels (i.e., 50, 500, and 1000 nM). The relative recovery was calculated using the following equation: Relative Recovery (%) =

Cspiked − Cnon Cadded

− spiked

× 100

where Cspiked and Cnon-spiked are the concentrations in the spiked and non-spiked samples, respectively, and Cadded is the spiked concentration of a known amount of standard. The concentration data in the relative recovery formula was calculated from the calibration

curve (Y = aX + b), and the Y-value was obtained after derivatization and AS-DLLME. The relative recoveries were 100–104%, as shown in Table 3. Because l-carnitine and its related compounds are endogenous, the standard addition method was used to quantify human plasma. Five different amounts of analytes in the standard solution were added to the plasma. Then, linear regression was performed based on the addition amount and analysis data (Table 4). The precision was measured by calculating the relative standard deviations (RSDs) in the slope and intercept in sextuplicate, and the RSDs were better than 7.50% and 9.06%, respectively. The labeling efficiency was evaluated using nano LC–MS/MS to compare the amounts of butyrobetaine, l-carnitine and acetyl-lcarnitine in the standard before and after derivatization. Because these analytes are endogenous substances, the experiment could not be performed using human plasma. The labeling efficiency was calculated by dividing the residual standard concentration by the original standard concentration followed by multiplication by 100. The labeling efficiencies of butyrobetaine, l-carnitine and acetyl-l-carnitine were 98.12%, 97.48% and 98.74%, respectively. The extraction efficiencies in standard and human plasma were compared in terms of the formation of butyrobetaine, l-carnitine and acetyl-l-carnitine derivatives after evaporation of the labeled samples (in ACN) followed by reconstitution in water or AS-DLLME. When the peak area for the simple evaporation was 100%, the extraction efficiencies of l-carnitine, butyrobetaine and acetyl-lcarnitine derivatives were 96.37%, 86.77% and 84.62%, respectively, in the standard solution and 99.08%, 88.62% and 78.50%, respectively, in human plasma. When the hydrophobicity of the analyte derivatives increased, the order of the extraction efficiency from highest to lowest was l-carnitine, butyrobetaine and acetyl-lcarnitine derivatives. Fig. 4 compares the chromatograms for the standard solution and plasma for simple evaporation followed by reconstitution in water (red line) or AS-DLLME (black line). The chromatograms indicate that the AS-DLLME resulted in clean chromatograms by preventing interference from the sample matrix. The stability of the butyrobetaine, l-carnitine and acetyl-lcarnitine derivatives was also investigated at room temperature for 12h after derivatization. The decrease in the fluorescence was less than 5% (n = 3) in the standard solution and 8% in the plasma samples, which indicated that the derivatives were sufficiently stable for narrow-bore LC-FLD analysis during this time period. 3.4. Applications 3.4.1. Analysis of pharmaceuticals, personal care products and food The proposed method was used to determine l-carnitine and acetyl-l-carnitine in 14 real samples. Figs. 2(b)–(d) show the chromatograms for these samples. Table 5 lists the determined amounts and percentages of the samples with respect to the content stated on the label. For pharmaceuticals A and B, the percentages of the samples with respect to the stated content were 99.2% and 102.1%, respectively. Notably, the United States Pharmacopeia requires the

Please cite this article in press as: Y.-C. Chen, et al., Fluorescent derivatization combined with aqueous solvent-based dispersive liquidliquid microextraction for determination of butyrobetaine, l-carnitine and acetyl-l-carnitine in human plasma, J. Chromatogr. A (2016), http://dx.doi.org/10.1016/j.chroma.2016.08.030

G Model

ARTICLE IN PRESS

CHROMA-357830; No. of Pages 10

Y.-C. Chen et al. / J. Chromatogr. A xxx (2016) xxx–xxx

7

Table 5 The l-carnitine and acetyl-l-carnitine assay results for pharmaceuticals, personal care products and food. Sample

Formulation

Amount found (␮g/g) l-Carnitine

Pharmaceutical A Pharmaceutical B Food A Food B Food C Food D Personal care product A Personal care product B Personal care product C Personal care product D Personal care product E Personal care product F Personal care product G Personal care product H a

Tablet Injection Capsule Candy Nutrition drink Coffee Gel Gel Shampoo Cream Lotion Gel Gel Cream

RSD (%)

Percentage of claimed content (%)

1.18 1.29 7.87 2.91 1.93 1.53 6.68 1.44 3.23 1.13 3.28 3.56 0.52 1.65

99.2 102.1 88.3 98.9 109.7

Acetyl-l-carnitine

9.9 × 102 a 2.0 × 102 a 4.4 × 102 a 2.5 × 102 a 8.8 × 101 6.6 × 10-1 2.9 × 103 8.9 × 10-1 3.3 × 102 1.5 × 104 4.9 × 102 4.9 × 102 4.8 × 102 1.7 × 102

Amount found unit is mg per g. Table 6 Analyte concentrations in plasma samples from healthy volunteers. Subject

1 2 3

Concentration (␮M) Butyrobetaine

l-Carnitine

Acetyl-l-carnitine

1.3 × 10 1.2 × 100 1.1 × 100

4.6 × 10 4.3 × 101 4.8 × 101

6.3 × 100 2.8 × 100 6.7 × 100

0

1

experimental procedure and calculation resulted in 1.1 × 100 –1.3 × 100 ␮M for butyrobetaine, 4.3 × 101 –4.8 × 101 ␮M for l-carnitine, and 2.8 × 100 –6.7 × 100 ␮M for acetyl-l-carnitine (Table 6). These results are consistent with those from previous studies, which reported butyrobetaine levels of 0.50–1.80 ␮M [19,38], l-carnitine levels of 17.50–66.14 ␮M [19,39] and acetyll-carnitine levels of 1.94–18.85 ␮M [19,38–40]. In healthy adults, the ratio of acyl-l-carnitine to l-carnitine is typically less than 0.25 [41]. Higher ratios indicate an abnormal metabolism. Nevertheless, acetyl-l-carnitine can be used to estimate acyl-l-carnitine because it is the most abundant acyl-l-carnitine. The acetyl-l-carnitine to l-carnitine ratios in all of the healthy subjects were in the normal range (i.e., 0.06–0.14). 3.5. Nano LC–MS/MS results

Fig. 4. Chromatograms for determining formation of butyrobetaine, l-carnitine and acetyl-l-carnitine derivatives by simple evaporation followed by reconstitution in water (red line) or with AS-DLLME (black line) in (A) standard solution and (B) plasma. Peak 1: l-carnitine derivative; peak 2: butyrobetaine derivative; peak 3: acetyl-l-carnitine. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

actual content of tablets and injections to be at least 90.0% and no more than 110.0% of the labeled content. The percentages of the other samples with respect to the stated content ranged from 88.3% to 109.7%. However, personal care products and food are not regulated. In unmarked samples of various formulations, l-carnitine and acetyl-l-carnitine were detected in concentrations ranging from 6.6 × 10−1 to 1.5 × 104 ␮g g−1 . 3.4.2. Analysis of human plasma Blood samples were collected from three randomly selected healthy adult volunteers. The value obtained by dividing the intercept by the slope of the regression equation was used to represent the original content of the sample in this study. The

These extremely hydrophilic analytes typically wash out in a reverse-phase column without retention in the LC–MS/MS system. Fluorescent derivatization combined with AS-DLLME is also applicable in LC–MS/MS. Nano LC–MS/MS was used for further confirmation of the structures of the derivatives. The AS-DLLME can extract acyl-l-carnitine derivatives other than butyrobetaine, l-carnitine, and acetyl-l-carnitine. Therefore, thirteen compounds related to l-carnitine were analyzed, and 11 compounds were short-, medium-, and long-chain acyl-l-carnitines (Table 7). The precursor ions were [M]+ , and the major product ions were located at m/z 167 or 226. The masses corresponded to the product ions shown in Fig. 5. These results indicate that the proposed derivatization method and AS-DLLME method enabled derivatization and detection of acyl-l-carnitines with widely varying polarities. 3.6. Comparison to other methods Table 8 compares the analytical parameters for determining l-carnitine and its related compounds by HPLC-FLD or by HPLC–MS/MS. Conventional methods require 15–500 ␮L of biosamples for the determination of l-carnitine and its related compounds. However, the current method only requires 1 ␮L of plasma

Please cite this article in press as: Y.-C. Chen, et al., Fluorescent derivatization combined with aqueous solvent-based dispersive liquidliquid microextraction for determination of butyrobetaine, l-carnitine and acetyl-l-carnitine in human plasma, J. Chromatogr. A (2016), http://dx.doi.org/10.1016/j.chroma.2016.08.030

G Model

ARTICLE IN PRESS

CHROMA-357830; No. of Pages 10

Y.-C. Chen et al. / J. Chromatogr. A xxx (2016) xxx–xxx

8

Table 7 Accuracy of mass measurements of butyrobetaine, l-carnitine and acyl-l-carnitines derivatives in human plasma detected by LC–MS/MS. Analyte

Formula [M]+

Exact mass (Da)

Measured mass (Da)

Butyrobetaine l-Carnitine Acetyl-l-carnitine Propionyl-l-carnitine Butyryl-l-carnitine Isovaleryl-l-carnitine Hexanoyl-l-carnitine Octanoyl-l-carnitine Decanoyl-l-carnitine Lauroyl-l-carnitine Myristoyl-l-carnitine Palmitoyl-l-carnitine Stearoyl-l-carnitine

C20 H26 NO2 C20 H26 NO3 C22 H28 NO4 C23 H30 NO4 C24 H32 NO4 C25 H34 NO4 C26 H36 NO4 C28 H40 NO4 C30 H44 NO4 C32 H48 NO4 C34 H52 NO4 C36 H56 NO4 C38 H60 NO4

312.1958 328.1907 370.2012 384.2175 398.2331 412.2487 426.2644 454.2957 482.3270 510.3583 538.3891 566.4204 594.4517

312.1955 328.1908 370.2008 384.2165 398.2325 412.2479 426.2638 454.2954 482.3265 510.3559 538.3884 566.4203 594.4516

Mass error (ppm) −0.96 0.30 −1.08 −2.60 −1.50 −1.94 −1.41 −0.66 −1.03 −4.70 −1.30 −0.18 −0.17

MS2 (m/z) 152, 167, 226 152, 167, 226, 310 152, 167, 226 152, 167, 226, 366 152, 167, 226, 380 152, 167, 226, 394 152, 167, 226, 408 152, 167, 226, 436 152, 167, 226, 464 152, 167, 226, 492 152, 167, 226, 520 152, 167, 226, 548 152, 167, 226, 576

Fig. 5. The MS2 general fragmentation pattern for butyrobetaine, l-carnitine and acyl-l-carnitine derivatives in human plasma. Table 8 Comparison of the proposed method and other methods described in the literature in terms of effectiveness for determination of l-carnitine and related compounds. Linear range (␮M)

Sample

Refs.

200 10 230

5–200 0.1–4 0–50

Serum

[12] [23]

–

–

Dried blood spots Plasma

100 100 –

0–228 0–228 30–130 /80–180

Plasma

[29]

Plasma/Urine

[17]

A-C

–

0–5.0

Plasma/Tissue

[18]

C AC C AC C

5000 1000 103000 19400 2000

5–160 1–32 100–1000 20–200 0−1000

Plasma

[19]

Urine

[20]

C B AC

3/−/3/-

0.01−2/−/0.01−2/-

Analytical method

Sample volume (␮L)

Derivatization Reaction reagent time (min)

Extraction method

Organic solvent volume (␮L)

Injection volume (␮L)

LOD (nM)

HPLC– MS/MS HPLC– MS/MS HPLC– MS/MS

20

Methanol

15

–

1100

1–2

15.2

Butanol

15

SE

485

–

C B AC

50

Butanol

20

–

700

2

HPLC– MS/MS HPLC-FLD

20

Butanol

–

SE

750

20

15

6 Mnatriflate 2 Panatriflate NaEtriflate

5

–

600

50

C AC A-C C AC t-C

10

SPE

9330

2

1-AA

20

SPE

10600

20

HPLC-FLD HPLC-FLD

50 /10–40a 100

HPLC-FLD

100–500

Br-DMEQ

60

SPE

4300

10

HPLC-FLD

50

p-BPB

90

–

1050

–

HPLC-FLD

20/1

Br-MBP

7/9

AS-DLLME

112/195

[28]

Seminal [21] plasma Pharmaceuticals, This work cosmetics, and food/Plasma

Abbreviation: 6 -Mnatriflate, 6 -methoxynaphthacyl triflate; 2 -Panatriflate, 2 -phenanthrenacyl triflate; 1-AA, 1-aminoanthracene; Br-DMEQ, 3-bromometbyl-6,7-methoxyl-methyl-2(lH)-quinoxalinone; p-BPB, p-bromophenacyl bromide; Br-MBP, 4-bromomethyl biphenyl; SE, solvent extraction; SPE, solid phase extraction; AS-DLLME, aqueous solvent-base dispersive liquid liquid microextraction; C, l-carnitine; B, butyrobetaine; AC, acetyl-l-carnitine; A-C, acyl-l-carnitines; t-C, total l-carnitine. a Milligram.

Please cite this article in press as: Y.-C. Chen, et al., Fluorescent derivatization combined with aqueous solvent-based dispersive liquidliquid microextraction for determination of butyrobetaine, l-carnitine and acetyl-l-carnitine in human plasma, J. Chromatogr. A (2016), http://dx.doi.org/10.1016/j.chroma.2016.08.030

G Model CHROMA-357830; No. of Pages 10

ARTICLE IN PRESS Y.-C. Chen et al. / J. Chromatogr. A xxx (2016) xxx–xxx

and is applicable for the analysis of pharmaceuticals, personal care products and food products. To the best of our knowledge, this study is the first to combine chemical derivatization with ASDLLME to determine l-carnitine in various samples. Derivatization improves the detection sensitivity, and AS-DLLME enhances the extractability of quaternary amines from biosamples. The proposed method is also environmentally friendly because it consumes only 195 ␮L of organic solvent. Its detection sensitivity is equal to or exceeds that of HPLC–MS/MS, which is the most commonly used method for clinical diagnosis. 4. Conclusions This study developed a simple and green methodology for detecting and extracting highly hydrophilic compounds from aqueous samples and biosamples via derivatization with AS-DLLME coupled with LC-FLD. The method was successfully used to monitor butyrobetaine, l-carnitine, and acetyl-l-carnitine in pharmaceuticals, personal care products, food products and human plasma. In terms of sample volume, consumption of organic solvent, reaction time and detection limit, the proposed method is comparable to or superior to both fluorescence detection and MS detection. Therefore, the developed approach is environmentally friendly and can be performed quickly and easily. Experimental applications of the method exhibited high sensitivity, precision and recovery for the extraction of aqueous samples under the optimized conditions. Therefore, AS-DLLME combined with the proposed method is effective for analyzing water-soluble compounds. Acknowledgments This project was supported by grant no. MOST 104-2113-M037-009 and MOST 105-2113-M-037-015 from the Ministry of Science and Technology and by Aim for the Top Universities grant no.KMU-TP104PR07 from Kaohsiung Medical University. Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.chroma.2016.08. 030. References [1] M. Rezaee, Y. Assadi, M.-R. MilaniHosseini, E. Aghaee, F. Ahmadi, S. Berijani, Determination of organic compounds in water using dispersive liquid-iquid microextraction, J. Chromatogr. A 1116 (2006) 1–9. [2] L. Guo, H.K. Lee, Vortex-assisted micro-solid-phase extraction followed by low-density solvent based dispersive liquid-liquid microextraction for the fast and efficient determination of phthalate esters in river water samples, J. Chromatogr. A 1300 (2013) 24–30. [3] M.A. Farajzadeh, D. Djozan, P. Khorram, Development of a new dispersive liquid-liquid microextraction method in a narrow-bore tube for preconcentration of triazole pesticides from aqueous samples, Anal. Chim. Acta 713 (2012) 70–78. [4] Y. Zhang, H.K. Lee, Low-density solvent-based vortex-assisted surfactant-enhanced-emulsification liquid-liquid microextraction combined with gas chromatography-mass spectrometry for the fast determination of phthalate esters in bottled water, J. Chromatogr. A 1274 (2013) 28–35. [5] A. Saleh, Y. Yamini, M. Faraji, M. Rezaee, M. Ghambarian, Ultrasound-assisted emulsification microextraction method based on applying low density organic solvents followed by gas chromatography analysis for the determination of polycyclic aromatic hydrocarbons in water samples, J. Chromatogr. A 1216 (2009) 6673–6679. [6] Z.H. Yang, P. Wang, W.T. Zhao, Z.Q. Zhou, D.H. Liu, Development of a home-made extraction device for vortex-assisted surfactant-enhanced-emulsification liquid-liquid microextraction with lighter than water organic solvents, J. Chromatogr. A 1300 (2013) 58–63. [7] L. Guo, S. Tan, X. Li, H.K. Lee, Fast automated dual-syringe based dispersive liquid–liquid microextraction coupled with gas chromatography–mass spectrometry for the determination of polycyclic aromatic hydrocarbons in environmental water samples, J. Chromatogr. A 1438 (2016) 1–9.

9

[8] B. Kamm, M. Kamm, A. Kiener, H.P. Meyer, Polycarnitine-a new biomaterial, Appl. Microbiol. Biotechnol. 67 (2005) 1–7. [9] C.J. Rebouche, D.J. Paulson, Carnitine metabolism and function in humans, Annual review of nutrition, Annu. Rev. Nutr. 6 (1986) 41–66. [10] K. Strijbis, F.M. Vaz, B. Distel, Enzymology of the carnitine biosynthesis pathway, IUBMB Life 62 (2010) 357–362. [11] J. Liu, H. Atamna, H. Kuratsune, B.N. Ames, Delaying brain mitochondrial decay and aging with mitochondrial antioxidants and metabolites, Ann. N. Y. Acad. Sci. 959 (2002) 133–166. [12] S. Friedman, Determination of carnitine in biological materials, Arch. Biochem. Biophys. 75 (1958) 24–30. [13] M.A. Mehlman, G. Wolf, Studies on the distribution of free carnitine and the occurrence and nature of bound carnitine, Arch. Biochem. Biophys. 98 (1962) 146–153. [14] G. Cederblad, S. Lindstedt, A method for the determination of carnitine in the picomole range, Clin. Chim. Acta 37 (1972) 235–243. [15] L. Vernez, W. Thormann, S. Krahenbuhl, Analysis of carnitine and acylcarnitines in urine by capillary electrophoresis, J. Chromatogr. A 895 (2000) 309–316. [16] W. Pormsila, S. Krahenbuhl, P.C. Hauser, Determination of carnitine in food and food supplements by capillary electrophoresis with contactless conductivity detection, Electrophoresis 31 (2010) 2186–2191. [17] L. Sánchez-Hernández, M. Castro-Puyana, C. García-Ruiz, A.L. Crego, M.L. Marina, Determination of l- and d-carnitine in dietary food supplements using capillary electrophoresis–tandem mass spectrometry, Food Chem. 120 (2010) 921–928. [18] L.M. Lewin, A. Peshin, B. Sklarz, A gas chromatographic assay for carnitine, Anal. Biochem. 68 (1975) 531–536. [19] F. Inoue, N. Terada, H. Nakajima, M. Okochi, N. Kodo, Z. Kizaki, A. Kinugasa, T. Sawada, Effect of sports activity on carnitine metabolism. Measurement of free carnitine, gamma-butyrobetaine and acylcarnitines by tandem mass spectrometry, J. Chromatogr. B 731 (1999) 83–88. [20] G.-X. He, T. Dahl, Improved high-performance liquid chromatographic method for analysis of l-carnitine in pharmaceutical formulations, J. Pharm. Biomed. 23 (2000) 315–321. [21] S. Krahenbuhl, P.E. Minkler, C.L. Hoppel, Derivatization of isolated endogenous butyrobetaine with 4 -bromophenacyl trifluoromethanesulfonate followed by high-performance liquid chromatography, J. Chromatogr. B 573 (1992) 3–10. [22] P.E. Minkler, C.L. Hoppel, Determination of free carnitine and ‘total’ carnitine in human urine: derivatization with 4 -bromophenacyl trifluoromethanesulfonate and high performance liquid chromatography, Clin. Chim. Acta 212 (1992) 55–64. [23] P.E. Minkler, E.P. Brass, W.R. Hiatt, S.T. Ingalls, C.L. Hoppel, Quantification of carnitine, acetylcarnitine, and total carnitine in tissues by high-performance liquid chromatography: the effect of exercise on carnitine homeostasis in man, Anal. Biochem. 231 (1995) 315–322. [24] C.J. McEntyre, M. Lever, M.K. Storer, A high performance liquid chromatographic method for the measurement of total carnitine in human plasma and urine, Clin. Chim. Acta 344 (2004) 123–130. [25] P.E. Minkler, J. Kerner, K.N. North, C.L. Hoppel, Quantitation of long-chain acylcarnitines by HPLC/fluorescence detection: application to plasma and tissue specimens from patients with carnitinepalmitoyltransferase-II deficiency, Clin. Chim. Acta 352 (2005) 81–92. [26] A. Longo, G. Bruno, S. Curti, A. Mancinelli, G. Miotto, Determination of l-carnitine, acetyl-l-carnitine and propionyl-l-carnitine in human plasma by high-performance liquid chromatography after pre-column derivatization with 1-aminoanthracene, J. Chromatogr. B 686 (1996) 129–139. [27] H. Kamimori, Y. Hamashima, M. Konishi, Determination of carnitine and saturated-acyl group carnitines in human urine by high-performance liquid chromatography with fluorescence detection, Anal. Biochem. 218 (1994) 417–424. [28] K. Li, W. Li, Y. Huang, Determination of free l-carnitine in human seminal plasma by high performance liquid chromatography with pre-column ultraviolet derivatization and its clinical application in male infertility, Clin. Chim. Acta 378 (2007) 159–163. [29] N. Kodo, D.S. Millington, D.L. Norwood, C.R. Roe, Quantitative assay of free and total carnitine using tandem mass spectrometry, Clin. Chim. Acta 186 (1989) 383–390. [30] D.H. Chace, S.L. Hillman, J.L. Van Hove, E.W. Naylor, Rapid diagnosis of MCAD deficiency: quantitative analysis of octanoylcarnitine and other acylcarnitines in newborn blood spots by tandem mass spectrometry, Clin. Chem. 43 (1997) 2106–2113. [31] C. Turgeon, M.J. Magera, P. Allard, S. Tortorelli, D. Gavrilov, D. Oglesbee, K. Raymond, P. Rinaldo, D. Matern, Combined newborn screening for succinylacetone amino acids, and acylcarnitines in dried blood spots, Clin. Chem. 54 (2008) 657–664. [32] M.G. Schooneman, N. Achterkamp, C.A. Argmann, M.R. Soeters, S.M. Houten, Plasma acylcarnitines inadequately reflect tissue acylcarnitine metabolism, Biochim. Biophys. Acta 1841 (2014) 987–994. [33] V.R. De Jesús, D.H. Chace, T.H. Lim, J.V. Mei, W.H. Hannon, Comparison of amino acids and acylcarnitines assay methods used in newborn screening assays by tandem mass spectrometry, Clin. Chim. Acta 411 (2010) 684–689. [34] J. Rasmussen, O.W. Nielsen, N. Janzen, M. Duno, L. Kober, U. Steuerwald, A.M. Lund, Carnitine levels in 26,462 individuals from the nationwide screening program for primary carnitine deficiency in the Faroe Islands, J. Inherit. Metab. Dis. 37 (2014) 215–222.

Please cite this article in press as: Y.-C. Chen, et al., Fluorescent derivatization combined with aqueous solvent-based dispersive liquidliquid microextraction for determination of butyrobetaine, l-carnitine and acetyl-l-carnitine in human plasma, J. Chromatogr. A (2016), http://dx.doi.org/10.1016/j.chroma.2016.08.030

G Model CHROMA-357830; No. of Pages 10 10

ARTICLE IN PRESS Y.-C. Chen et al. / J. Chromatogr. A xxx (2016) xxx–xxx

[35] I. Ferrer, P. Ruiz-Sala, Y. Vicente, B. Merinero, C. Pérez-Cerdá, M. Ugarte, Separation and identification of plasma short-chain acylcarnitine isomers by HPLC/MS/MS for the differential diagnosis of fatty acid oxidation defects and organic acidemias, J. Chromatogr. B 860 (2007) 121–126. [36] A. Liu, M. Pasquali, Acidified acetonitrile and methanol extractions for quantitative analysis of acylcarnitines in plasma by stable isotope dilution tandem mass spectrometry, J. Chromatogr. B 827 (2005) 193–198. [37] J. McMurry, Carbonyl condensation reactions, in: L. Lockwood (Ed.), Organic Chemistry, Cengage Learning, Belmont, 2010, pp. 904–908. [38] F. Hirche, M. Fischer, J. Keller, K. Eder, Determination of carnitine, its short chain acyl esters and metabolic precursors trimethyllysine and gamma-butyrobetaine by quasi-solid phase extraction and MS/MS detection, J. Chromatogr. B 877 (2009) 2158–2162.

[39] M. Peng, X. Fang, Y. Huang, Y. Cai, C. Liang, R. Lin, L. Liu, Separation and identification of underivatized plasma acylcarnitine isomers using liquid chromatography-tandem mass spectrometry for the differential diagnosis of organic acidemias and fatty acid oxidation defects, J. Chromatogr. A 1319 (2013) 97–106. [40] E. De Palo, R. Gatti, M. Varnier, A. Floreani, C. De Palo, C. Scandellari1, Plasma acetyl-carnitine concentrations during and after a muscular exercise test in patients with liver disease, Eur. J. Clin. Chem. Clin. Biochem. 30 (1992) 179–186. [41] D.K. Hothi, D.F. Geary, L. Fisher, C.T. Chan, Short-term effects of nocturnal haemodialysis on carnitine metabolism, Nephrol. Dial. Transplant. 21 (2006) 2637–2641.

Please cite this article in press as: Y.-C. Chen, et al., Fluorescent derivatization combined with aqueous solvent-based dispersive liquidliquid microextraction for determination of butyrobetaine, l-carnitine and acetyl-l-carnitine in human plasma, J. Chromatogr. A (2016), http://dx.doi.org/10.1016/j.chroma.2016.08.030