- Email: [email protected]
Nonaqueous capillary electrophoresis with laser-induced fluorescence detection: A case study of comparison with aqueous media
a n a l y t i c a c h i m i c a a c t a 6 1 1 ( 2 0 0 8 ) 212–219
available at www.sciencedirect.com
journal homepage: www.elsevier.com/locate/aca
Nonaqueous capillary electrophoresis with laser-induced fluorescence detection: A case study of comparison with aqueous media Lei Zhou a , Weiping Wang a , Shumin Wang a , Yang Hui a , Zhi Luo a , Zhide Hu a,b,∗ a b
Department of Chemistry, Lanzhou University, Lanzhou 730000, China Lanzhou Institute of Separation Science, Lanzhou 730070, China
a r t i c l e
i n f o
a b s t r a c t
Article history:
A novel method based on separation by nonaqueous capillary electrophoresis (NACE) com-
Received 29 November 2007
bined with laser-induced fluorescence (LIF) detection was developed and compared with
Received in revised form
classic aqueous modes of electrophoresis in terms of resolution of solutes of interest and
25 January 2008
sensitivity of the fluorescence detection. Catecholamines derivatized with 4-chloro-7-nitro-
Accepted 28 January 2008
2,1,3-benzoxadiazole (NBD-Cl) were chosen as test analytes for their subtle fluorescence
Published on line 13 February 2008
properties. In aqueous systems, capillary zone electrophoresis (CZE) was not suitable for the analysis of test analytes due to complete fluorescence quenching of NBD-labeled cat-
Keywords:
echolamines in neat aqueous buffer. The addition of micelles or microemulsion droplets
Aprotic solvent derivatization
into aqueous running buffer can dramatically improve the fluorescence response, and the
Nonaqueous capillary
enhancement seems to be comparable for micellar electrokinetic chromatography (MEKC)
electrophoresis
and microemulsion electrokinetic chromatography (MEEKC). As another alternative, NACE
Laser-induced fluorescence
separation was advantageous when performing the analysis under the optimum separa-
Catecholamine
tion condition of 20 mM sodium tetraborate, 20 mM sodium dodecyl sulfate (SDS), 0.1% (v/v)
4-Chloro-7-nitro-2,1,3-
glacial acetic acid, 20% (v/v) acetonitrile (ACN) in methanol medium after derivatization in
benzoxadiazole
ACN/dimethyl sulfoxide (DMSO) (3:2, v/v) mixed aprotic solvents containing 20 mM ammonium acetate. Compared with derivatization and separation in aqueous media, NACE–LIF procedure was proved to be superior, providing high sensitivity and short migration time. Under respective optimum conditions, the NACE procedure offered the best fluorescence response with 5–24 folds enhancement for catecholamines compared to aqueous procedures. In addition, the mechanisms of derivatization and separation in nonaqueous media were elucidated in detail. © 2008 Elsevier B.V. All rights reserved.
1.
Introduction
Capillary electrophoresis (CE) in conjunction with laserinduced fluorescence detection, with its highly efficient resolving power and utmost sensitivity, has documented a useful technique for the analysis of a wide number of
∗
compounds, especially for bioanalytical applications such as DNA, proteins and single cell analysis [1–3]. CE–LIF, however, often presents problems with fluorescence quenching by quenchers in buffer systems such as molecular oxygen (O2 ) or impurities [4,5]. Among many intermolecular or intramolecular processes responsible for fluorescence quenching [6],
Corresponding author at: Department of Chemistry, Lanzhou University, Lanzhou 730000, China. Fax: +86 931 891 2582. E-mail addresses: [email protected] (L. Zhou), [email protected] (Z. Hu). 0003-2670/$ – see front matter © 2008 Elsevier B.V. All rights reserved. doi:10.1016/j.aca.2008.01.084
a n a l y t i c a c h i m i c a a c t a 6 1 1 ( 2 0 0 8 ) 212–219
two main mechanisms, dynamic quenching and long-range energy transfer, have been proposed to account for quenching at low concentration of quencher in buffer systems [4,6]. Dynamic quenching occurs when an excited-state molecule collides with quenching species producing a non-radiative energy transfer. This process is diffusion-controlled and thus dependent on the temperature and viscosity of the medium. Long-range energy transfer also occurs due to a non-radiative energy transfer as a result of dipole interactions between the fluorophore and quencher without a collision. Generally, fluorescence quenching is a confused problem with fluorescence detection, which leads to decreased signal output and results in poor detection limits. Addition of micelles or microemulsion droplets in the buffer is an effective attempt to minimize the quenching effects in CE–LIF methods [7,8]. It is well known that fluorescence intensity strongly depends on the microenvironment around the fluorophore. The presence of micelles or microemulsion in background electrolyte would make the fluorophore incorporate into them, and then (i) the collision is attenuated as a result of the shielding effect; (ii) the O2 diffusion rate is decreased due to being more viscous microenvironment; (iii) other deactivation and quenching processes are reduced in a more organized microenvironment [4,9]. For instance, it has been reported that micelles have been utilized to minimize quenching for the determination of amino acids [10] and peptides [11,12]. Our group has reported similar results using microemulsion droplets as additives in background electrolyte for quantitative determination of amino acid derivatives [13] and ephedrine and pseudoephedrine [14]. In addition, it is reported that complexation with cyclodextrins can be employed for an improvement of the sensitivity of fluorescence detection [15], which is based on an enhancement of the fluorescence quantum yield due to a decrease in the rotation motion of the entrapped molecule and/or decrease in solvent relaxation. Other attempts at minimizing the quenching effects involve the introduction of organic solvents into the buffers [16]. It is well known that the fluorescent properties are strongly dependent on the nature of solvents such as dielectric constant, polarity and viscosity [17–19]. The use of organic solvents can alter the viscosity and polarity of the medium. By making the medium more viscous and less polar, the dynamic quenching and long-range energy transfer can be minimized. Khaledi and co-worker [4] were the first to describe the use of NACE with LIF to improve detection sensitivity. And they found that nonaqueous solvent, N-methylformamide, produced the best detection limit with a two-fold enhancement using NACE–LIF as compared to aqueous CE. Couderc and co-workers [5] reported the detection of nitrotyrosine and tyrosine derivatived with 4-fluoro-7nitro-2,1,3-benzoxadiazole (NBD-F), and found that only in organic media the peaks of derivatives can be detected, though the separation is not obvious. However, the mechanisms of derivatization and separation in nonaqueous media were not elucidated clearly in aforecited researches, and the related study is still very scarce. Compounds with the benzofurazan skeleton, such as NBDCl and NBD-F, are commonly used as fluorogenic reagent for amine derivatization [20]. This kind of fluorogenic reagent
213
is useful for its mild reaction conditions and good matching with convenient argon-ion LIF detector. However, it is weakly fluorescent in aqueous buffer [2]. In our previous work for simultaneous determination of catecholamines and amino acids by MEKC-LIF [21], we found that no signal was observed for NBD-Cl labeled catecholamines using a conventional aqueous buffer even containing 10% (v/v) ACN. Then we introduced SDS micelles into the running buffer to analyze the same derivatives and the results indicated that the introduction of micelles had a predominant effect on the fluorescence response. As a continuing research, in this case study, we investigated and compared the influence of different derivatization and separation behaviors in aqueous and nonaqueous media on the fluorescence intensity using NBD-labeled catecholamines as test analytes by CE with LIF detection. Compared to derivatization and separation in aqueous media, a better result with higher sensitivity and shorter migration time can be achieved in nonaqueous media. Also, the mechanisms of derivatization and separation in nonaqueous media were elucidated in detail.
2.
Experimental
2.1.
Chemicals
Catecholamines norepinephrine (NE), epinephrine (E), and dopamine (DA) were obtained from Sigma (Steinheim, Germany). NBD-Cl was the product of Acros Organics (Geel, Belgium). SDS was supplied by Shanghai Shiyi Chemicals Reagent Factory (Shanghai, China). All other chemicals were of analytical reagent grade and were used as received. Distilled water was used throughout the study. Stock standard solutions with NE (100 g mL−1 ), E (50 g mL−1 ) and DA (100 g mL−1 ) were prepared by dissolving the appropriate amount compounds in distilled water (containing 2.5 mM HCl) for aqueous medium derivatization and in methanol (containing 2.5 mM HCl) for nonaqueous medium derivatization, respectively. The required working solutions were obtained by appropriate dilution of stock solution with distilled water or methanol. One hundred millimolar stock solution of NBD-Cl was prepared in ACN and diluted to the desired concentration with ACN before use. All the stock solutions were stored at 4 ◦ C.
2.2.
Electrolyte solutions
2.2.1.
Aqueous electrolyte solutions
A mixed electrolyte solution consisted of 5 mM sodium tetraborate, 80 mM boric acid and 20 mM potassium chloride (KCl) was used for derivatization in aqueous medium and adjusted to the desired pH with 0.5 M HCl or NaOH solution. Running buffers for different separation modes were prepared from stock solution of 100 mM sodium tetraborate, ACN, stock solution of 200 mM SDS or stock microemulsion solution, and then adjusted to the desired pH using 0.5 M HCl or NaOH solution. A stock microemulsion solution composed of 3.0% (v/v) ethyl acetate/150 mM SDS/3.0% (v/v) 1-butanol, was prepared by addition of 3 mL ethyl acetate, 4.3257 g SDS and 3 mL 1butanol to a 100 mL volumetric flask and diluted to 100 mL
214
a n a l y t i c a c h i m i c a a c t a 6 1 1 ( 2 0 0 8 ) 212–219
Table 1 – Comparison of the optimum derivatization and separation conditions in aqueous and nonaqueous media Medium Aqueous
Derivatization conditions 5 mM sodium tetraborate, 80 mM boric acid and 20 mM KCl buffer at pH 8.0, 30 mM NBD-Cl, reaction at 60 ◦ C for 20 min
Separation mode CZE MEKC
MEEKC
Nonaqueous
20 mM ammonium acetate in the ACN/DMSO mixed aprotic solvent with the ratio of 3:2 (v/v), 30 mM NBD-Cl, and reaction at 60 ◦ C for 20 min
with distilled water. The mixture was sonicated for 20 min to aid dissolution and then an optically transparent and stable microemulsion was formed. Before use, all buffer solutions were filtered through a 0.45 m cellulose acetate membrane and degassed in an ultrasonic bath.
2.2.2.
Nonaqueous electrolyte solutions
The derivatization buffer, 20 mM ammonium acetate in ACN/DMSO (3:2, v/v) mixed solvents, was prepared daily from ACN and stock solution of 200 mM ammonium acetate in DMSO. The methanol-based running buffer, typically consisted of 20 mM sodium tetraborate, 20 mM SDS, 20% (v/v) ACN and 0.1% (v/v) glacial acetic acid, was freshly prepared from SDS, ACN and stock solution of 100 mM sodium tetraborate in methanol and then adjusted by addition of glacial acetic acid or triethylamine to the required acidity. All nonaqueous electrolyte solutions were filtered through a 0.45 m polyvinylidene fluoride membrane and degassed in an ultrasonic bath prior to use.
2.3.
Derivatization procedures
The derivatization procedure in aqueous medium was performed as previously described [21]. A similar procedure for nonaqueous medium derivatization was carried out as follows: 20 L of stock standard solution, 100 L of the derivatization buffer, and 100 L of NBD-Cl solution was added sequentially into a 500 L microcentrifuge tube. After thoroughly mixed, derivatization reaction was conducted at 60 ◦ C for 20 min in a water bath under dark. Then, 180 L of derivatization buffer was added into the reaction vial, and the total volume of the mixture reached up to 400 L. After cooling to room temperature, the mixture was introduced into the capillary by hydrodynamic injection without further dilution. For blank solution, it was prepared like that of standard solutions.
2.4.
Apparatus and electrophoretic conditions
All separations were performed on a Beckman P/ACETM 5510 system (Fullerton, CA, USA) equipped with a LIF detector. The excitation light from an argon-ion laser (3 mW) was focused on the capillary window by means of a fiber-optic connection. Excitation was performed at 488 nm and the elec-
NACE
Separation conditions
Typical figure
10 mM sodium tetraborate, 20% (v/v) ACN at pH 9.9 10 mM sodium tetraborate, 25 mM SDS, 20% (v/v) ACN at pH 9.9 0.6% (v/v) ethyl acetate, 30 mM SDS, 0.6% (v/v) 1-butanol, 22% (v/v) ACN, 10 mM sodium tetraborate at pH 9.9 20 mM sodium tetraborate, 20 mM SDS, 0.1% (v/v) glacial acetic acid, 20% (v/v) ACN in methanol medium
Fig. 1a Fig. 1b
Fig. 1c
Fig. 2b
tropherograms were recorded by monitoring the emission intensity at 520 nm. Data were collected and analyzed by means of Beckman P/ACETM station software. Fused-silica capillary with dimensions of 50 m I.D. (375 m O.D.) × 47 cm (40 cm to the detection window) was purchased from Yongnian Ruifeng Chromatogram Equipment (Yongnian, China), which was accommodated in a cartridge configured for LIF detection. A new capillary was preconditioned prior to use by rinsing sequentially with methanol for 5 min, distilled water for 2 min, 1.0 M HCl for 5 min, distilled water for 2 min, 0.1 M NaOH for 15 min, distilled water for 2 min, running buffer for 15 min and, finally, equilibrated at 20 kV with running buffer for 20 min. The capillary was thermostated at 25 ◦ C with a liquid coolant system. For all separation modes, samples were introduced from the anodic end of the capillary by applying a pressure of 3.45 kPa for 3 s and separations were conducted with a voltage of 25 kV. The variables for different separation modes were systematically investigated and the final optimum systems were summarized in Table 1. Additional experimental details are given in the legends to tables and figures. All buffers were renewed after every three runs to obtain good repeatability. In addition, the repeatability of migration time was affected by rinsing procedures between runs [22] due to the pH hysteresis effect with silica capillary [23]. As a result, different rinsing procedures between runs were adopted based on different separation modes: the rinsing procedure for MEKC involved rinsing with distilled water for 2 min, 0.1 M HCl for 2 min, distilled water for 2 min, 0.1 M NaOH for 2 min, distilled water for 2 min, running buffer for 2 min; for MEEKC involved rinsing with distilled water for 2 min, 0.1 M NaOH for 2 min, distilled water for 2 min, 0.1 M HCl for 2 min, distilled water for 2 min, running buffer for 2 min; while for NACE, only rinsing with running buffer for 2 min was involved.
3.
Results and discussion
3.1. Derivatization and separation in aqueous media Investigation of the derivatization conditions in aqueous medium for three catecholamines labeled with NBD-Cl was
a n a l y t i c a c h i m i c a a c t a 6 1 1 ( 2 0 0 8 ) 212–219
215
3.2. Derivatization and separation in nonaqueous media 3.2.1. Investigation of the nonaqueous derivatization conditions
Fig. 1 – Typical electropherograms of CZE, MEKC and MEEKC for the standard solution at the concentration level of 5.00, 5.00 and 2.50 g mL−1 for NE, DA and E, respectively. Derivatization conditions: 5 mM sodium tetraborate, 80 mM boric acid and 20 mM KCl buffer at pH 8.0, 30 mM NBD-Cl, reaction at 60 ◦ C for 20 min. Separation conditions: CZE: 10 mM sodium tetraborate, 20% (v/v) ACN at pH 9.9; MEKC: 10 mM sodium tetraborate, 25 mM SDS, 20% (v/v) ACN at pH 9.9; MEEKC: 0.6% (v/v) ethyl acetate/30 mM SDS/0.6% (v/v) 1-butanol/22% (v/v) ACN/10 mM sodium tetraborate at pH 9.9. Peak identification: 1-NE, 2-DA, 3-E, unidentified peaks correspond to the hydrolysis products of NBD-Cl.
achieved by testing fluorescence signal intensity. The effects of buffer pH, reaction temperature, reaction time, and NBD-Cl concentration were investigated systematically and the optimum result was listed in Table 1, similar to that reported in our previous work [21]. For all aqueous modes of electrophoresis, the same derivatization conditions were applied. Once the derivatization conditions in aqueous medium had been investigated, the subsequent investigation of separation conditions was performed including three aqueous separation modes: CZE, MEKC and MEEKC. In CZE, it was found that except for the hydrolysis peak of NBD-Cl, no peaks were eluted under all tested conditions, even the introduction of 20% (v/v) ACN into the running buffer (Fig. 1a). This may be attributed to that the fluorescence of NBDlabeled catecholamines was quenched completely. It is worthy pointing out that an attempt by applying other electrolytes such as phosphate, acetate or Tris had also been performed in the preliminary experiments and the results were similar to that of tetraborate. Then, micelles or microemulsion droplets were introduced into the running buffer, as shown in Fig. 1b and c, the results were obvious that the fluorescence response of the derivatives was observed. This is explained that addition of micelles or microemulsion droplets altered the microenviroment of buffer solution, thereby minimizing the fluorescence quenching effects. Subsequently, the effects of MEKC and MEEKC parameters on fluorescence response intensity and separation efficiency were investigated, and consequent separation conditions were summarized in Table 1.
When the samples of catecholamines derivatived in aqueous media were analyzed directly by NACE, the fluorescence responses presented a little improvement. However, a puzzled problem, current breakdown, was encountered in the course of separation as a result of the dismatch between aqueous sample zone and nonaqueous running buffer. Thus, it is essential to develop a novel derivatization method in nonaqueous media for NACE–LIF. Recently, Dong et al. [24] reported the derivatization of hydrophobic amino acids and ephedrine and pseudoephedrine [25] in methanol medium and separation by NACE with LIF detection. But that is a failed attempt due to poor sensitivity and long separation time. In this study, we had also attempted primarily to carry out the derivatization reaction in methanol-based buffer, but the fluorescence responses were too weak to be observed (Fig. 2a). This may be ascribed to that methanol is considered as protic solvent, unfavorable to nucleophilic reaction. ACN, a dipolar aprotic solvent, was then chosen as the background solvent for nonaqueous derivatization buffer. Unfortunately, ammonium acetate, a common electrolyte in nonaqueous media, cannot be dissolved completely in ACN. So another aprotic solvent DMSO was introduced to enhance the solubility. Interestingly enough, when the aprotic solvent system of ACN/DMSO was used as nonaqueous derivatization buffer, a remarkable improved fluorescence response was observed, as shown in Fig. 2b. Then, the variables affecting nonaqueous derivatization efficiency, the ratio of ACN to DMSO, electrolyte concentration, buffer acidity, reaction temperature and time, were investigated. The ratio of ACN to DMSO has a profound effect
Fig. 2 – Comparison of the derivatization in nonaqueous media. Derivatization conditions: 20 mM ammonium acetate in methanol (a) or in ACN/DMSO (3:2, v/v) mixed solvent (b), 30 mM NBD-Cl, reaction at 60 ◦ C for 20 min. Separation conditions: 20 mM sodium tetraborate, 20 mM SDS, 0.1% (v/v) glacial acetic acid, 20% (v/v) ACN in methanol medium. Peaks identification as in Fig. 1.
216
a n a l y t i c a c h i m i c a a c t a 6 1 1 ( 2 0 0 8 ) 212–219
on the fluorescence responses of catecholamine derivatives. The fluorescence responses increased as the DMSO proportion increased from 20 to 40% (v/v), and further increase above 40% was followed by a decrease in the fluorescence response. Consequently, a value of 40% (v/v) DMSO, i.e. the ratio of ACN to DMSO was 3:2, was applied for further studies. Ammonium acetate, as the electrolyte for derivatization buffer, was employed in this work and investigated in the range of 10–40 mM. It was found that 20 mM ammonium acetate can provide the satisfactory results. Buffer acidity, as another dependent variable affecting the derivatization efficiency, was examined by addition of acetic acid (0–0.4%, v/v) or triethylamine (0–0.2%, v/v). The results implied that an addition of acetic acid or triethylamine did not contribute to the increase in the fluorescence response, but caused an opposite result. Hence, the optimum acidity for nonaqueous derivatization was obtained without addition of acetic acid or triethylamine. In addition, other derivatization variables, reaction temperature and time, as well as NBD-Cl concentration were also investigated and the results were similar to that in aqueous derivatization. Therefore, the optimum derivatization conditions in nonaqueous medium were chosen as follows: 20 mM ammonium acetate in the ACN/DMSO mixed aprotic solvent with the ratio of 3:2 (v/v), 30 mM NBD-Cl, and reaction at 60 ◦ C for 20 min.
3.2.2. Investigation of the nonaqueous separation conditions The fluorescence properties of the purified NBD derivatives have been studied in a range of organic solvents using fluorescence spectra [17–19], and the results demonstrated that the fluorescent intensity and emission maximum wavelength were strongly dependent on the nature of solvents such as dielectric constant, polarity and viscosity. Considering the different requirements for the successful use of NACE, the physicochemical properties of organic solvents are vital to control electroosmotic flow (EOF) and analyte migration [26,27]. Methanol and acetonitrile, as the most common solvents in NACE, were therefore recommended in this study. Electrolyte is described also an important factor influencing the separation and sensitivity in NACE. In order to select an appropriate electrolyte, we studied the effects of the nature and concentration of different electrolytes (ammonium acetate, Tris and sodium tetraborate) on separation behavior using the same initial condition of 20 mM SDS, 10% (v/v) ACN, and 0.48% (v/v) triethylamine in methanol medium. The comparison of separation profiles is shown in Fig. 3, suggesting that sodium tetraborate can provide the best resolution and peak shape for all analytes, most likely due to the complexation of borate with the cis-diol of catecholamine molecules [28–30]. This complexation between borate and catecholamines results in more negative charge promoting the negative species to migrate towards the anode. Moreover, the different stability of catecholamine complex with borate enlarges the difference in mobility of these analytes, which is beneficial to improve the resolution. Therefore, sodium tetraborate was chosen as the electrolyte in nonaqueous media. On the other hand, information derived from Fig. 3 indicate that the fluorescence signal intensity of derivatives are similar for three different electrolyte systems, suggesting that elec-
Fig. 3 – Effect of different electrolytes on NACE separation; (a) 20 mM sodium tetraborate, (b) 20 mM Tris and (c) 20 mM ammonium acetate, other separation conditions: 20 mM SDS, 0.48% (v/v) triethylamine, 10% (v/v) ACN in methanol medium. Derivatization conditions: 20 mM ammonium acetate in ACN/DMSO (4:1, v/v) mixed solvent, 30 mM NBD-Cl, reaction at 60 ◦ C for 20 min. Peaks identification as in Fig. 1.
trolyte has little or no fluorescence quenching effect on the NBD derivatives of catecholamines. The results are in accordance with those obtained from fluorescence spectroscopic studies [17]. The presence of ionic additives, SDS in this work, has also vital influence on separation selectivity and resolution, as shown in Fig. 4. With an increase in SDS concentration, a noteworthy increase in separation selectivity and resolution can be achieved. As we know, the critical micelle concentration (CMC) of SDS in aqueous media approximates 8 mM. If micelles can
Fig. 4 – Effect of SDS concentration on the migration time and run current in NACE system. Derivatization conditions: 20 mM ammonium acetate in ACN/DMSO (3:2, v/v) mixed solvent, 30 mM NBD-Cl, reaction at 60 ◦ C for 20 min. Separation conditions: 20 mM sodium tetraborate, 0.1% (v/v) glacial acetic acid, 20% (v/v) ACN in methanol medium.
217
a n a l y t i c a c h i m i c a a c t a 6 1 1 ( 2 0 0 8 ) 212–219
be formed in the nonaqueous methanol solution, the CMC will be much higher than in aqueous media because methanol is less polar than water [31]. In this study, micelles cannot be formed in the nonaqueous methanol solution even SDS concentration reaches 25 mM. This is also proved by means of the method proposed by Cifuentes et al. [32]. In Fig. 4, by plotting the electric current values, as measured by a CE instrument, versus the SDS concentration, a good linearity has been obtained with the relative coefficient of 0.9991, supporting that micelles are absent within the range of investigations in nonaqueous methanol solution. Nevertheless, our results indicate that SDS monomers still can interact with analytes to form ion pairs, which is favored to enhance electrophoretic mobility differences and further improve separation selectivity [33,34]. The employed condition of 20 mM SDS appears to offer an acceptable balance between the resolution and migration time. The acidity of the electrolyte is another governing factor on the separation behavior and fluorescence response. In this paper, the electrolyte acidity was adjusted by addition of glacial acetic acid or triethylamine. The addition of glacial acetic acid could provide higher fluorescence response and shorter migration time, companied with a decreased resolution; whereas the introduction of triethylamine would result in lower fluorescence response and longer migration time, though the resolution was increased. Hence, 0.1% (v/v) glacial acetic acid was recommended as a compromise. In addition, an incorporation of ACN into methanol medium was found particularly advantageous to achieve high selectivity and short migration time. In this study, the effect of ACN concentration was investigated in the range of 0–30% (v/v), and 20% (v/v) ACN offered the optimum result. According to the factors mentioned above, the optimum separation was obtained with 20 mM sodium tetraborate, 20 mM SDS, 0.1% (v/v) glacial acetic acid, 20% (v/v) ACN in methanol medium and an applied voltage of 25 kV (Fig. 2b).
3.3.
Comparative study
To evaluate the separation behaviors in different separation modes (CZE, MEKC, MEEKC and NACE), a standard solution containing 5.00 g mL−1 NE, 5.00 g mL−1 DA and 2.50 g mL−1 E was analyzed under respective optimum conditions (Table 1) as shown in Figs. 1 and 2b. A comparison of peak area for three analytes at the same concentration was exhibited in Fig. 5. It was clear that no signal was observed using CZE mode; and for MEKC and MEEKC systems, the separation profiles were quite similar in fluorescence response, but a little longer migration time with the latter. Compared to separation in aqueous media, NACE offered the potential for higher fluorescence responses and shorter migration times. It can be seen from Fig. 5 that, NACE–LIF produced the best fluorescence response with a 24-, 5-, and 5-fold enhancement for NE, E and DA relative to MEEKC. Furthermore, six consecutive replica runs were performed to evaluate the repeatability for migration time and peak area under respective optimum conditions, and the relative standard deviations (RSDs) were summarized in Table 2. The data implied that compared to the aqueous systems, inferior repeatability for migration time and peak area was obtained in nonaqueous media due to the volatility of
Fig. 5 – Comparison of fluorescence responses for NBD-labeled catecholamines in optimum CZE, MEKC, MEEKC and NACE systems. The standard solution was used at the concentration level of 5.00, 5.00 and 2.50 g mL−1 for NE, DA and E, respectively.
the organic electrolyte system. Then, the standard solutions of catecholamines in the concentration rang of 0.10–5.00 g mL−1 were used to determine the analytical characteristics of different separation modes. Figures of merit were listed in Table 3, suggesting that high linearity was attainable for all separation modes. The slope of a calibration curve is often used as a measure of analytical sensitivity. The slopes of different separation modes have the following order: NACE > MEKC ≈ MEEKC. The higher slope of the nonaqueous calibration curves is an indication of sensitivity enhancement. The limit of detection (LOD) in different systems was calculated based on a signal-to-noise ratio of 3, showing the following order: NACE < MEKC ≈ MEEKC. It can be seen from Table 3 that NACE–LIF produced the best sensitivity with 3–6 folds enhancement for catecholamines as compared to aqueous systems. However, compared to other derivatization agents such as fluorescein isothiocyanate (FITC) [35,36], the sensitivity of the present NACE–LIF method was not satisfactory as a result of inherent weak fluorescence of NBD-Cl. So the NACE–LIF method was not applied for real samples in the present work. Nevertheless, these discussions on derivatization and separation behaviors in nonaqueous media were significant and guidable for further study.
Table 2 – Comparison of repeatability for migration time and peak area in MEKC, MEEKC and NACE systems (RSD, %, n = 6) Speciesa
Migration time MEKC
NE E DA a
0.29 0.53 0.45
MEEKC 0.46 0.49 0.56
NACE 0.87 0.64 0.61
Peak area MEKC 2.9 2.8 3.1
MEEKC
NACE
3.5 4.0 5.0
5.8 4.6 4.8
The standard solution was used in repeatability study at the concentration level of 5.00 g mL−1 for NE and DA, 2.50 g mL−1 for E.
218
a n a l y t i c a c h i m i c a a c t a 6 1 1 ( 2 0 0 8 ) 212–219
Table 3 – Comparison of analytical characteristics in MEKC, MEEKC and NACE systems Species
MEKC Slope 2.98 × 10 1.55 × 106 6.41 × 105 5
NE E DA a
4.
r 0.9999 0.9994 0.9978
MEEKC DLa
Slope
72.6 14.0 33.8
2.88 × 10 1.56 × 106 6.88 × 105
r 5
0.9996 0.9991 0.9993
NACE DL
Slope
80.5 14.8 33.7
1.54 × 10 4.41 × 106 2.36 × 106 6
r
DL
0.9940 0.9949 0.9919
12.4 4.3 8.1
Detection limit (ng mL−1 ) based on a signal/noise ratio of 3.
Conclusions
In the present work, we developed a novel NACE–LIF method using NBD-labeled catecholamines as test analytes and presented the method comparison in selectivity and sensitivity with classic aqueous modes of electrophoresis including CZE, MEKC and MEEKC. In aqueous systems, MEKC and MEEKC are conducive in method development when coupled with LIF detection for that the addition of micelles or microemulsion droplets into running buffer could efficiently minimize the fluorescence quenching. However, NACE–LIF was found to be superior to aqueous systems in terms of separation selectivity and detection sensitivity. Additionally, the NACE–LIF method offered much faster analysis time and simple rinsing procedure between runs, therefore demonstrating a promising technique. Importantly, the effect of solvents on nonaqueous derivatization reaction was demonstrated, and a mixed aprotic solvent (ACN/DMSO) was favorable to nucleophilic reaction in nonaqueous media. In NACE separation, the high resolution and short migration time was obtained by using sodium tetraborate as electrolyte in methanol medium, which was likely attributed to the complexation effect of borate with catecholamines. Furthermore, the separation selectivity and resolution of catecholamines derivatives was further improved by adding SDS as ionic additives to methanol electrolyte. Our results indicate that even though micelles are not present in operating buffer, SDS monomers still interact with analytes to form ion pairs, which is favored to enhance mobility differences and to improve analyte separation. The authors hope that these discussions on derivatization and separation behaviors in nonaqueous media will significantly extend the application area of NACE with LIF detection.
Acknowledgement The authors express their sincere gratitude to Dr. Zhiyong Tang for helpful discussions about nonaqueous solvent chemistry.
references
[1] J. Kraly, M.A. Fazal, R.M. Schoenherr, R. Bonn, M.M. Harwood, E. Turner, M. Jones, N.J. Dovichi, Anal. Chem. 78 (2006) 4097. [2] M. Lacroix, V. Poinsot, C. Fournier, F. Couderc, Electrophoresis 26 (2005) 2608.
[3] Y.-W. Lin, T.-C. Chiu, H.-T. Chang, J. Chromatogr. B 793 (2003) 37. [4] V.L. Ward, M.G. Khaledi, J. Chromatogr. B 718 (1998) 15. [5] C. Massip, P. Riollet, E. Quemener, C. Bayle, R. Salvayre, F. ´ J. Chromatogr. A 979 (2002) 209. Couderc, E. Causse, [6] B. Valeur, Molecular Fluorescence: Principles and Applications, Wiley-VCH, Weinheim, 2001. [7] M. Molina, M. Silva, Electrophoresis 23 (2002) 3907. [8] A. Marsh, B. Clark, M. Broderick, J. Power, S. Donegan, K. Altria, Electrophoresis 25 (2004) 3970. [9] G.G. Guilbault (Ed.), Practical Fluorescence, Marcel Dekker, New York, 1990. [10] S. Hu, P.C.H. Li, J. Chromatogr. A 876 (2000) 183. [11] I. Beijersten, D. Westerlund, J. Chromatogr. A 716 (1995) 389. [12] S. Descroix, I.L. Potier, C. Niquet, N. Minc, J. Viovy, M. Taverna, J. Chromatogr. A 1087 (2005) 203. [13] J. Xie, J. Zhang, H. Liu, J. Liu, J. Tian, X. Chen, Z. Hu, Biomed. Chromatogr. 18 (2004) 600. [14] J. Zhang, J. Xie, J. Liu, J. Tian, X. Chen, Z. Hu, Electrophoresis 25 (2004) 74. ¨ [15] C. Huhn, M. Putz, N. Martin, R. Dahlenburg, U. Pyell, Electrophoresis 26 (2005) 2391. [16] E.M.R. Rodr´ıguez, M.S. Alaejos, C.D. Romero, Anal. Chim. Acta 334 (1996) 161. [17] X. Zhu, P.N. Shaw, D.A. Barrett, Anal. Chim. Acta 478 (2003) 259. [18] J.A. Monti, S.T. Christian, W.A. Shaw, J. Lipid Res. 19 (1978) 222. [19] M. Bem, F. Badea, C. Draghici, M.T. Caproiu, M. Vasilescu, M. Voicescu, A. Beteringhe, A. Caragheorgheopol, M. Maganu, T. Constantinescu, A.T. Balaban, ARKIVOC xiii (2007) 87. [20] S. Uchiyama, T. Santa, N. Okiyama, T. Fukushima, K. Imai, Biomed. Chromatogr. 15 (2001) 295. [21] L. Zhou, S. Wang, K. Tian, Y. Dong, Z. Hu, J. Sep. Sci. 30 (2007) 110. ¨ [22] H. Watzig, M. Degenhardt, A. Kunkel, Electrophoresis 19 (1998) 2695. [23] W.J. Lambert, D.L. Middleton, Anal. Chem. 62 (1990) 1585. [24] Y. Dong, X. Chen, Z. Hu, Microchim. Acta 154 (2006) 281. [25] Y. Dong, Y. Chen, X. Chen, Z. Hu, Biomed. Chromatogr. 20 (2006) 1150. ´ J. Chromatogr. [26] M.L. Riekkola, M. Jussila, S.P. Porras, I.E. Valko, A 892 (2000) 155. [27] F. Steiner, M. Hassel, Electrophoresis 21 (2000) 3994. [28] S. Hoffstetter-Kuhn, A. Paulus, E. Gassmann, H.M. Widmer, Anal. Chem. 63 (1991) 1541. [29] J.P. Landers, R.P. Oda, M.D. Schuchard, Anal. Chem. 64 (1992) 2846. [30] M. Stefansson, M. Novotny, J. Am. Chem. Soc. 115 (1993) 11573. [31] W. Lu, G.K. Poon, P.L. Carmichael, R.B. Cole, Anal. Chem. 68 (1996) 668. [32] A. Cifuentes, J.L. Bernal, J.C. Diez-Masa, Anal. Chem. 69 (1997) 4271.
a n a l y t i c a c h i m i c a a c t a 6 1 1 ( 2 0 0 8 ) 212–219
[33] R.M. Seifar, J.C. Kraak, W.T. Kok, Anal. Chem. 69 (1997) 2772. [34] R. Carabias-Mart´ınez, E. Rodr´ıguez-Gonzalo, J. ´ ´ Dom´ınguez-Alvarez, J. Hernandez-M endez, Anal. Chem. 69 (1997) 4437.
219
[35] J. Zhang, X. Chen, Z. Hu, X. Ma, Anal. Chim. Acta 471 (2002) 203. [36] M. Du, V. Flanigan, Y. Ma, Electrophoresis 25 (2004) 1496.
available at www.sciencedirect.com
journal homepage: www.elsevier.com/locate/aca
Nonaqueous capillary electrophoresis with laser-induced fluorescence detection: A case study of comparison with aqueous media Lei Zhou a , Weiping Wang a , Shumin Wang a , Yang Hui a , Zhi Luo a , Zhide Hu a,b,∗ a b
Department of Chemistry, Lanzhou University, Lanzhou 730000, China Lanzhou Institute of Separation Science, Lanzhou 730070, China
a r t i c l e
i n f o
a b s t r a c t
Article history:
A novel method based on separation by nonaqueous capillary electrophoresis (NACE) com-
Received 29 November 2007
bined with laser-induced fluorescence (LIF) detection was developed and compared with
Received in revised form
classic aqueous modes of electrophoresis in terms of resolution of solutes of interest and
25 January 2008
sensitivity of the fluorescence detection. Catecholamines derivatized with 4-chloro-7-nitro-
Accepted 28 January 2008
2,1,3-benzoxadiazole (NBD-Cl) were chosen as test analytes for their subtle fluorescence
Published on line 13 February 2008
properties. In aqueous systems, capillary zone electrophoresis (CZE) was not suitable for the analysis of test analytes due to complete fluorescence quenching of NBD-labeled cat-
Keywords:
echolamines in neat aqueous buffer. The addition of micelles or microemulsion droplets
Aprotic solvent derivatization
into aqueous running buffer can dramatically improve the fluorescence response, and the
Nonaqueous capillary
enhancement seems to be comparable for micellar electrokinetic chromatography (MEKC)
electrophoresis
and microemulsion electrokinetic chromatography (MEEKC). As another alternative, NACE
Laser-induced fluorescence
separation was advantageous when performing the analysis under the optimum separa-
Catecholamine
tion condition of 20 mM sodium tetraborate, 20 mM sodium dodecyl sulfate (SDS), 0.1% (v/v)
4-Chloro-7-nitro-2,1,3-
glacial acetic acid, 20% (v/v) acetonitrile (ACN) in methanol medium after derivatization in
benzoxadiazole
ACN/dimethyl sulfoxide (DMSO) (3:2, v/v) mixed aprotic solvents containing 20 mM ammonium acetate. Compared with derivatization and separation in aqueous media, NACE–LIF procedure was proved to be superior, providing high sensitivity and short migration time. Under respective optimum conditions, the NACE procedure offered the best fluorescence response with 5–24 folds enhancement for catecholamines compared to aqueous procedures. In addition, the mechanisms of derivatization and separation in nonaqueous media were elucidated in detail. © 2008 Elsevier B.V. All rights reserved.
1.
Introduction
Capillary electrophoresis (CE) in conjunction with laserinduced fluorescence detection, with its highly efficient resolving power and utmost sensitivity, has documented a useful technique for the analysis of a wide number of
∗
compounds, especially for bioanalytical applications such as DNA, proteins and single cell analysis [1–3]. CE–LIF, however, often presents problems with fluorescence quenching by quenchers in buffer systems such as molecular oxygen (O2 ) or impurities [4,5]. Among many intermolecular or intramolecular processes responsible for fluorescence quenching [6],
Corresponding author at: Department of Chemistry, Lanzhou University, Lanzhou 730000, China. Fax: +86 931 891 2582. E-mail addresses: [email protected] (L. Zhou), [email protected] (Z. Hu). 0003-2670/$ – see front matter © 2008 Elsevier B.V. All rights reserved. doi:10.1016/j.aca.2008.01.084
a n a l y t i c a c h i m i c a a c t a 6 1 1 ( 2 0 0 8 ) 212–219
two main mechanisms, dynamic quenching and long-range energy transfer, have been proposed to account for quenching at low concentration of quencher in buffer systems [4,6]. Dynamic quenching occurs when an excited-state molecule collides with quenching species producing a non-radiative energy transfer. This process is diffusion-controlled and thus dependent on the temperature and viscosity of the medium. Long-range energy transfer also occurs due to a non-radiative energy transfer as a result of dipole interactions between the fluorophore and quencher without a collision. Generally, fluorescence quenching is a confused problem with fluorescence detection, which leads to decreased signal output and results in poor detection limits. Addition of micelles or microemulsion droplets in the buffer is an effective attempt to minimize the quenching effects in CE–LIF methods [7,8]. It is well known that fluorescence intensity strongly depends on the microenvironment around the fluorophore. The presence of micelles or microemulsion in background electrolyte would make the fluorophore incorporate into them, and then (i) the collision is attenuated as a result of the shielding effect; (ii) the O2 diffusion rate is decreased due to being more viscous microenvironment; (iii) other deactivation and quenching processes are reduced in a more organized microenvironment [4,9]. For instance, it has been reported that micelles have been utilized to minimize quenching for the determination of amino acids [10] and peptides [11,12]. Our group has reported similar results using microemulsion droplets as additives in background electrolyte for quantitative determination of amino acid derivatives [13] and ephedrine and pseudoephedrine [14]. In addition, it is reported that complexation with cyclodextrins can be employed for an improvement of the sensitivity of fluorescence detection [15], which is based on an enhancement of the fluorescence quantum yield due to a decrease in the rotation motion of the entrapped molecule and/or decrease in solvent relaxation. Other attempts at minimizing the quenching effects involve the introduction of organic solvents into the buffers [16]. It is well known that the fluorescent properties are strongly dependent on the nature of solvents such as dielectric constant, polarity and viscosity [17–19]. The use of organic solvents can alter the viscosity and polarity of the medium. By making the medium more viscous and less polar, the dynamic quenching and long-range energy transfer can be minimized. Khaledi and co-worker [4] were the first to describe the use of NACE with LIF to improve detection sensitivity. And they found that nonaqueous solvent, N-methylformamide, produced the best detection limit with a two-fold enhancement using NACE–LIF as compared to aqueous CE. Couderc and co-workers [5] reported the detection of nitrotyrosine and tyrosine derivatived with 4-fluoro-7nitro-2,1,3-benzoxadiazole (NBD-F), and found that only in organic media the peaks of derivatives can be detected, though the separation is not obvious. However, the mechanisms of derivatization and separation in nonaqueous media were not elucidated clearly in aforecited researches, and the related study is still very scarce. Compounds with the benzofurazan skeleton, such as NBDCl and NBD-F, are commonly used as fluorogenic reagent for amine derivatization [20]. This kind of fluorogenic reagent
213
is useful for its mild reaction conditions and good matching with convenient argon-ion LIF detector. However, it is weakly fluorescent in aqueous buffer [2]. In our previous work for simultaneous determination of catecholamines and amino acids by MEKC-LIF [21], we found that no signal was observed for NBD-Cl labeled catecholamines using a conventional aqueous buffer even containing 10% (v/v) ACN. Then we introduced SDS micelles into the running buffer to analyze the same derivatives and the results indicated that the introduction of micelles had a predominant effect on the fluorescence response. As a continuing research, in this case study, we investigated and compared the influence of different derivatization and separation behaviors in aqueous and nonaqueous media on the fluorescence intensity using NBD-labeled catecholamines as test analytes by CE with LIF detection. Compared to derivatization and separation in aqueous media, a better result with higher sensitivity and shorter migration time can be achieved in nonaqueous media. Also, the mechanisms of derivatization and separation in nonaqueous media were elucidated in detail.
2.
Experimental
2.1.
Chemicals
Catecholamines norepinephrine (NE), epinephrine (E), and dopamine (DA) were obtained from Sigma (Steinheim, Germany). NBD-Cl was the product of Acros Organics (Geel, Belgium). SDS was supplied by Shanghai Shiyi Chemicals Reagent Factory (Shanghai, China). All other chemicals were of analytical reagent grade and were used as received. Distilled water was used throughout the study. Stock standard solutions with NE (100 g mL−1 ), E (50 g mL−1 ) and DA (100 g mL−1 ) were prepared by dissolving the appropriate amount compounds in distilled water (containing 2.5 mM HCl) for aqueous medium derivatization and in methanol (containing 2.5 mM HCl) for nonaqueous medium derivatization, respectively. The required working solutions were obtained by appropriate dilution of stock solution with distilled water or methanol. One hundred millimolar stock solution of NBD-Cl was prepared in ACN and diluted to the desired concentration with ACN before use. All the stock solutions were stored at 4 ◦ C.
2.2.
Electrolyte solutions
2.2.1.
Aqueous electrolyte solutions
A mixed electrolyte solution consisted of 5 mM sodium tetraborate, 80 mM boric acid and 20 mM potassium chloride (KCl) was used for derivatization in aqueous medium and adjusted to the desired pH with 0.5 M HCl or NaOH solution. Running buffers for different separation modes were prepared from stock solution of 100 mM sodium tetraborate, ACN, stock solution of 200 mM SDS or stock microemulsion solution, and then adjusted to the desired pH using 0.5 M HCl or NaOH solution. A stock microemulsion solution composed of 3.0% (v/v) ethyl acetate/150 mM SDS/3.0% (v/v) 1-butanol, was prepared by addition of 3 mL ethyl acetate, 4.3257 g SDS and 3 mL 1butanol to a 100 mL volumetric flask and diluted to 100 mL
214
a n a l y t i c a c h i m i c a a c t a 6 1 1 ( 2 0 0 8 ) 212–219
Table 1 – Comparison of the optimum derivatization and separation conditions in aqueous and nonaqueous media Medium Aqueous
Derivatization conditions 5 mM sodium tetraborate, 80 mM boric acid and 20 mM KCl buffer at pH 8.0, 30 mM NBD-Cl, reaction at 60 ◦ C for 20 min
Separation mode CZE MEKC
MEEKC
Nonaqueous
20 mM ammonium acetate in the ACN/DMSO mixed aprotic solvent with the ratio of 3:2 (v/v), 30 mM NBD-Cl, and reaction at 60 ◦ C for 20 min
with distilled water. The mixture was sonicated for 20 min to aid dissolution and then an optically transparent and stable microemulsion was formed. Before use, all buffer solutions were filtered through a 0.45 m cellulose acetate membrane and degassed in an ultrasonic bath.
2.2.2.
Nonaqueous electrolyte solutions
The derivatization buffer, 20 mM ammonium acetate in ACN/DMSO (3:2, v/v) mixed solvents, was prepared daily from ACN and stock solution of 200 mM ammonium acetate in DMSO. The methanol-based running buffer, typically consisted of 20 mM sodium tetraborate, 20 mM SDS, 20% (v/v) ACN and 0.1% (v/v) glacial acetic acid, was freshly prepared from SDS, ACN and stock solution of 100 mM sodium tetraborate in methanol and then adjusted by addition of glacial acetic acid or triethylamine to the required acidity. All nonaqueous electrolyte solutions were filtered through a 0.45 m polyvinylidene fluoride membrane and degassed in an ultrasonic bath prior to use.
2.3.
Derivatization procedures
The derivatization procedure in aqueous medium was performed as previously described [21]. A similar procedure for nonaqueous medium derivatization was carried out as follows: 20 L of stock standard solution, 100 L of the derivatization buffer, and 100 L of NBD-Cl solution was added sequentially into a 500 L microcentrifuge tube. After thoroughly mixed, derivatization reaction was conducted at 60 ◦ C for 20 min in a water bath under dark. Then, 180 L of derivatization buffer was added into the reaction vial, and the total volume of the mixture reached up to 400 L. After cooling to room temperature, the mixture was introduced into the capillary by hydrodynamic injection without further dilution. For blank solution, it was prepared like that of standard solutions.
2.4.
Apparatus and electrophoretic conditions
All separations were performed on a Beckman P/ACETM 5510 system (Fullerton, CA, USA) equipped with a LIF detector. The excitation light from an argon-ion laser (3 mW) was focused on the capillary window by means of a fiber-optic connection. Excitation was performed at 488 nm and the elec-
NACE
Separation conditions
Typical figure
10 mM sodium tetraborate, 20% (v/v) ACN at pH 9.9 10 mM sodium tetraborate, 25 mM SDS, 20% (v/v) ACN at pH 9.9 0.6% (v/v) ethyl acetate, 30 mM SDS, 0.6% (v/v) 1-butanol, 22% (v/v) ACN, 10 mM sodium tetraborate at pH 9.9 20 mM sodium tetraborate, 20 mM SDS, 0.1% (v/v) glacial acetic acid, 20% (v/v) ACN in methanol medium
Fig. 1a Fig. 1b
Fig. 1c
Fig. 2b
tropherograms were recorded by monitoring the emission intensity at 520 nm. Data were collected and analyzed by means of Beckman P/ACETM station software. Fused-silica capillary with dimensions of 50 m I.D. (375 m O.D.) × 47 cm (40 cm to the detection window) was purchased from Yongnian Ruifeng Chromatogram Equipment (Yongnian, China), which was accommodated in a cartridge configured for LIF detection. A new capillary was preconditioned prior to use by rinsing sequentially with methanol for 5 min, distilled water for 2 min, 1.0 M HCl for 5 min, distilled water for 2 min, 0.1 M NaOH for 15 min, distilled water for 2 min, running buffer for 15 min and, finally, equilibrated at 20 kV with running buffer for 20 min. The capillary was thermostated at 25 ◦ C with a liquid coolant system. For all separation modes, samples were introduced from the anodic end of the capillary by applying a pressure of 3.45 kPa for 3 s and separations were conducted with a voltage of 25 kV. The variables for different separation modes were systematically investigated and the final optimum systems were summarized in Table 1. Additional experimental details are given in the legends to tables and figures. All buffers were renewed after every three runs to obtain good repeatability. In addition, the repeatability of migration time was affected by rinsing procedures between runs [22] due to the pH hysteresis effect with silica capillary [23]. As a result, different rinsing procedures between runs were adopted based on different separation modes: the rinsing procedure for MEKC involved rinsing with distilled water for 2 min, 0.1 M HCl for 2 min, distilled water for 2 min, 0.1 M NaOH for 2 min, distilled water for 2 min, running buffer for 2 min; for MEEKC involved rinsing with distilled water for 2 min, 0.1 M NaOH for 2 min, distilled water for 2 min, 0.1 M HCl for 2 min, distilled water for 2 min, running buffer for 2 min; while for NACE, only rinsing with running buffer for 2 min was involved.
3.
Results and discussion
3.1. Derivatization and separation in aqueous media Investigation of the derivatization conditions in aqueous medium for three catecholamines labeled with NBD-Cl was
a n a l y t i c a c h i m i c a a c t a 6 1 1 ( 2 0 0 8 ) 212–219
215
3.2. Derivatization and separation in nonaqueous media 3.2.1. Investigation of the nonaqueous derivatization conditions
Fig. 1 – Typical electropherograms of CZE, MEKC and MEEKC for the standard solution at the concentration level of 5.00, 5.00 and 2.50 g mL−1 for NE, DA and E, respectively. Derivatization conditions: 5 mM sodium tetraborate, 80 mM boric acid and 20 mM KCl buffer at pH 8.0, 30 mM NBD-Cl, reaction at 60 ◦ C for 20 min. Separation conditions: CZE: 10 mM sodium tetraborate, 20% (v/v) ACN at pH 9.9; MEKC: 10 mM sodium tetraborate, 25 mM SDS, 20% (v/v) ACN at pH 9.9; MEEKC: 0.6% (v/v) ethyl acetate/30 mM SDS/0.6% (v/v) 1-butanol/22% (v/v) ACN/10 mM sodium tetraborate at pH 9.9. Peak identification: 1-NE, 2-DA, 3-E, unidentified peaks correspond to the hydrolysis products of NBD-Cl.
achieved by testing fluorescence signal intensity. The effects of buffer pH, reaction temperature, reaction time, and NBD-Cl concentration were investigated systematically and the optimum result was listed in Table 1, similar to that reported in our previous work [21]. For all aqueous modes of electrophoresis, the same derivatization conditions were applied. Once the derivatization conditions in aqueous medium had been investigated, the subsequent investigation of separation conditions was performed including three aqueous separation modes: CZE, MEKC and MEEKC. In CZE, it was found that except for the hydrolysis peak of NBD-Cl, no peaks were eluted under all tested conditions, even the introduction of 20% (v/v) ACN into the running buffer (Fig. 1a). This may be attributed to that the fluorescence of NBDlabeled catecholamines was quenched completely. It is worthy pointing out that an attempt by applying other electrolytes such as phosphate, acetate or Tris had also been performed in the preliminary experiments and the results were similar to that of tetraborate. Then, micelles or microemulsion droplets were introduced into the running buffer, as shown in Fig. 1b and c, the results were obvious that the fluorescence response of the derivatives was observed. This is explained that addition of micelles or microemulsion droplets altered the microenviroment of buffer solution, thereby minimizing the fluorescence quenching effects. Subsequently, the effects of MEKC and MEEKC parameters on fluorescence response intensity and separation efficiency were investigated, and consequent separation conditions were summarized in Table 1.
When the samples of catecholamines derivatived in aqueous media were analyzed directly by NACE, the fluorescence responses presented a little improvement. However, a puzzled problem, current breakdown, was encountered in the course of separation as a result of the dismatch between aqueous sample zone and nonaqueous running buffer. Thus, it is essential to develop a novel derivatization method in nonaqueous media for NACE–LIF. Recently, Dong et al. [24] reported the derivatization of hydrophobic amino acids and ephedrine and pseudoephedrine [25] in methanol medium and separation by NACE with LIF detection. But that is a failed attempt due to poor sensitivity and long separation time. In this study, we had also attempted primarily to carry out the derivatization reaction in methanol-based buffer, but the fluorescence responses were too weak to be observed (Fig. 2a). This may be ascribed to that methanol is considered as protic solvent, unfavorable to nucleophilic reaction. ACN, a dipolar aprotic solvent, was then chosen as the background solvent for nonaqueous derivatization buffer. Unfortunately, ammonium acetate, a common electrolyte in nonaqueous media, cannot be dissolved completely in ACN. So another aprotic solvent DMSO was introduced to enhance the solubility. Interestingly enough, when the aprotic solvent system of ACN/DMSO was used as nonaqueous derivatization buffer, a remarkable improved fluorescence response was observed, as shown in Fig. 2b. Then, the variables affecting nonaqueous derivatization efficiency, the ratio of ACN to DMSO, electrolyte concentration, buffer acidity, reaction temperature and time, were investigated. The ratio of ACN to DMSO has a profound effect
Fig. 2 – Comparison of the derivatization in nonaqueous media. Derivatization conditions: 20 mM ammonium acetate in methanol (a) or in ACN/DMSO (3:2, v/v) mixed solvent (b), 30 mM NBD-Cl, reaction at 60 ◦ C for 20 min. Separation conditions: 20 mM sodium tetraborate, 20 mM SDS, 0.1% (v/v) glacial acetic acid, 20% (v/v) ACN in methanol medium. Peaks identification as in Fig. 1.
216
a n a l y t i c a c h i m i c a a c t a 6 1 1 ( 2 0 0 8 ) 212–219
on the fluorescence responses of catecholamine derivatives. The fluorescence responses increased as the DMSO proportion increased from 20 to 40% (v/v), and further increase above 40% was followed by a decrease in the fluorescence response. Consequently, a value of 40% (v/v) DMSO, i.e. the ratio of ACN to DMSO was 3:2, was applied for further studies. Ammonium acetate, as the electrolyte for derivatization buffer, was employed in this work and investigated in the range of 10–40 mM. It was found that 20 mM ammonium acetate can provide the satisfactory results. Buffer acidity, as another dependent variable affecting the derivatization efficiency, was examined by addition of acetic acid (0–0.4%, v/v) or triethylamine (0–0.2%, v/v). The results implied that an addition of acetic acid or triethylamine did not contribute to the increase in the fluorescence response, but caused an opposite result. Hence, the optimum acidity for nonaqueous derivatization was obtained without addition of acetic acid or triethylamine. In addition, other derivatization variables, reaction temperature and time, as well as NBD-Cl concentration were also investigated and the results were similar to that in aqueous derivatization. Therefore, the optimum derivatization conditions in nonaqueous medium were chosen as follows: 20 mM ammonium acetate in the ACN/DMSO mixed aprotic solvent with the ratio of 3:2 (v/v), 30 mM NBD-Cl, and reaction at 60 ◦ C for 20 min.
3.2.2. Investigation of the nonaqueous separation conditions The fluorescence properties of the purified NBD derivatives have been studied in a range of organic solvents using fluorescence spectra [17–19], and the results demonstrated that the fluorescent intensity and emission maximum wavelength were strongly dependent on the nature of solvents such as dielectric constant, polarity and viscosity. Considering the different requirements for the successful use of NACE, the physicochemical properties of organic solvents are vital to control electroosmotic flow (EOF) and analyte migration [26,27]. Methanol and acetonitrile, as the most common solvents in NACE, were therefore recommended in this study. Electrolyte is described also an important factor influencing the separation and sensitivity in NACE. In order to select an appropriate electrolyte, we studied the effects of the nature and concentration of different electrolytes (ammonium acetate, Tris and sodium tetraborate) on separation behavior using the same initial condition of 20 mM SDS, 10% (v/v) ACN, and 0.48% (v/v) triethylamine in methanol medium. The comparison of separation profiles is shown in Fig. 3, suggesting that sodium tetraborate can provide the best resolution and peak shape for all analytes, most likely due to the complexation of borate with the cis-diol of catecholamine molecules [28–30]. This complexation between borate and catecholamines results in more negative charge promoting the negative species to migrate towards the anode. Moreover, the different stability of catecholamine complex with borate enlarges the difference in mobility of these analytes, which is beneficial to improve the resolution. Therefore, sodium tetraborate was chosen as the electrolyte in nonaqueous media. On the other hand, information derived from Fig. 3 indicate that the fluorescence signal intensity of derivatives are similar for three different electrolyte systems, suggesting that elec-
Fig. 3 – Effect of different electrolytes on NACE separation; (a) 20 mM sodium tetraborate, (b) 20 mM Tris and (c) 20 mM ammonium acetate, other separation conditions: 20 mM SDS, 0.48% (v/v) triethylamine, 10% (v/v) ACN in methanol medium. Derivatization conditions: 20 mM ammonium acetate in ACN/DMSO (4:1, v/v) mixed solvent, 30 mM NBD-Cl, reaction at 60 ◦ C for 20 min. Peaks identification as in Fig. 1.
trolyte has little or no fluorescence quenching effect on the NBD derivatives of catecholamines. The results are in accordance with those obtained from fluorescence spectroscopic studies [17]. The presence of ionic additives, SDS in this work, has also vital influence on separation selectivity and resolution, as shown in Fig. 4. With an increase in SDS concentration, a noteworthy increase in separation selectivity and resolution can be achieved. As we know, the critical micelle concentration (CMC) of SDS in aqueous media approximates 8 mM. If micelles can
Fig. 4 – Effect of SDS concentration on the migration time and run current in NACE system. Derivatization conditions: 20 mM ammonium acetate in ACN/DMSO (3:2, v/v) mixed solvent, 30 mM NBD-Cl, reaction at 60 ◦ C for 20 min. Separation conditions: 20 mM sodium tetraborate, 0.1% (v/v) glacial acetic acid, 20% (v/v) ACN in methanol medium.
217
a n a l y t i c a c h i m i c a a c t a 6 1 1 ( 2 0 0 8 ) 212–219
be formed in the nonaqueous methanol solution, the CMC will be much higher than in aqueous media because methanol is less polar than water [31]. In this study, micelles cannot be formed in the nonaqueous methanol solution even SDS concentration reaches 25 mM. This is also proved by means of the method proposed by Cifuentes et al. [32]. In Fig. 4, by plotting the electric current values, as measured by a CE instrument, versus the SDS concentration, a good linearity has been obtained with the relative coefficient of 0.9991, supporting that micelles are absent within the range of investigations in nonaqueous methanol solution. Nevertheless, our results indicate that SDS monomers still can interact with analytes to form ion pairs, which is favored to enhance electrophoretic mobility differences and further improve separation selectivity [33,34]. The employed condition of 20 mM SDS appears to offer an acceptable balance between the resolution and migration time. The acidity of the electrolyte is another governing factor on the separation behavior and fluorescence response. In this paper, the electrolyte acidity was adjusted by addition of glacial acetic acid or triethylamine. The addition of glacial acetic acid could provide higher fluorescence response and shorter migration time, companied with a decreased resolution; whereas the introduction of triethylamine would result in lower fluorescence response and longer migration time, though the resolution was increased. Hence, 0.1% (v/v) glacial acetic acid was recommended as a compromise. In addition, an incorporation of ACN into methanol medium was found particularly advantageous to achieve high selectivity and short migration time. In this study, the effect of ACN concentration was investigated in the range of 0–30% (v/v), and 20% (v/v) ACN offered the optimum result. According to the factors mentioned above, the optimum separation was obtained with 20 mM sodium tetraborate, 20 mM SDS, 0.1% (v/v) glacial acetic acid, 20% (v/v) ACN in methanol medium and an applied voltage of 25 kV (Fig. 2b).
3.3.
Comparative study
To evaluate the separation behaviors in different separation modes (CZE, MEKC, MEEKC and NACE), a standard solution containing 5.00 g mL−1 NE, 5.00 g mL−1 DA and 2.50 g mL−1 E was analyzed under respective optimum conditions (Table 1) as shown in Figs. 1 and 2b. A comparison of peak area for three analytes at the same concentration was exhibited in Fig. 5. It was clear that no signal was observed using CZE mode; and for MEKC and MEEKC systems, the separation profiles were quite similar in fluorescence response, but a little longer migration time with the latter. Compared to separation in aqueous media, NACE offered the potential for higher fluorescence responses and shorter migration times. It can be seen from Fig. 5 that, NACE–LIF produced the best fluorescence response with a 24-, 5-, and 5-fold enhancement for NE, E and DA relative to MEEKC. Furthermore, six consecutive replica runs were performed to evaluate the repeatability for migration time and peak area under respective optimum conditions, and the relative standard deviations (RSDs) were summarized in Table 2. The data implied that compared to the aqueous systems, inferior repeatability for migration time and peak area was obtained in nonaqueous media due to the volatility of
Fig. 5 – Comparison of fluorescence responses for NBD-labeled catecholamines in optimum CZE, MEKC, MEEKC and NACE systems. The standard solution was used at the concentration level of 5.00, 5.00 and 2.50 g mL−1 for NE, DA and E, respectively.
the organic electrolyte system. Then, the standard solutions of catecholamines in the concentration rang of 0.10–5.00 g mL−1 were used to determine the analytical characteristics of different separation modes. Figures of merit were listed in Table 3, suggesting that high linearity was attainable for all separation modes. The slope of a calibration curve is often used as a measure of analytical sensitivity. The slopes of different separation modes have the following order: NACE > MEKC ≈ MEEKC. The higher slope of the nonaqueous calibration curves is an indication of sensitivity enhancement. The limit of detection (LOD) in different systems was calculated based on a signal-to-noise ratio of 3, showing the following order: NACE < MEKC ≈ MEEKC. It can be seen from Table 3 that NACE–LIF produced the best sensitivity with 3–6 folds enhancement for catecholamines as compared to aqueous systems. However, compared to other derivatization agents such as fluorescein isothiocyanate (FITC) [35,36], the sensitivity of the present NACE–LIF method was not satisfactory as a result of inherent weak fluorescence of NBD-Cl. So the NACE–LIF method was not applied for real samples in the present work. Nevertheless, these discussions on derivatization and separation behaviors in nonaqueous media were significant and guidable for further study.
Table 2 – Comparison of repeatability for migration time and peak area in MEKC, MEEKC and NACE systems (RSD, %, n = 6) Speciesa
Migration time MEKC
NE E DA a
0.29 0.53 0.45
MEEKC 0.46 0.49 0.56
NACE 0.87 0.64 0.61
Peak area MEKC 2.9 2.8 3.1
MEEKC
NACE
3.5 4.0 5.0
5.8 4.6 4.8
The standard solution was used in repeatability study at the concentration level of 5.00 g mL−1 for NE and DA, 2.50 g mL−1 for E.
218
a n a l y t i c a c h i m i c a a c t a 6 1 1 ( 2 0 0 8 ) 212–219
Table 3 – Comparison of analytical characteristics in MEKC, MEEKC and NACE systems Species
MEKC Slope 2.98 × 10 1.55 × 106 6.41 × 105 5
NE E DA a
4.
r 0.9999 0.9994 0.9978
MEEKC DLa
Slope
72.6 14.0 33.8
2.88 × 10 1.56 × 106 6.88 × 105
r 5
0.9996 0.9991 0.9993
NACE DL
Slope
80.5 14.8 33.7
1.54 × 10 4.41 × 106 2.36 × 106 6
r
DL
0.9940 0.9949 0.9919
12.4 4.3 8.1
Detection limit (ng mL−1 ) based on a signal/noise ratio of 3.
Conclusions
In the present work, we developed a novel NACE–LIF method using NBD-labeled catecholamines as test analytes and presented the method comparison in selectivity and sensitivity with classic aqueous modes of electrophoresis including CZE, MEKC and MEEKC. In aqueous systems, MEKC and MEEKC are conducive in method development when coupled with LIF detection for that the addition of micelles or microemulsion droplets into running buffer could efficiently minimize the fluorescence quenching. However, NACE–LIF was found to be superior to aqueous systems in terms of separation selectivity and detection sensitivity. Additionally, the NACE–LIF method offered much faster analysis time and simple rinsing procedure between runs, therefore demonstrating a promising technique. Importantly, the effect of solvents on nonaqueous derivatization reaction was demonstrated, and a mixed aprotic solvent (ACN/DMSO) was favorable to nucleophilic reaction in nonaqueous media. In NACE separation, the high resolution and short migration time was obtained by using sodium tetraborate as electrolyte in methanol medium, which was likely attributed to the complexation effect of borate with catecholamines. Furthermore, the separation selectivity and resolution of catecholamines derivatives was further improved by adding SDS as ionic additives to methanol electrolyte. Our results indicate that even though micelles are not present in operating buffer, SDS monomers still interact with analytes to form ion pairs, which is favored to enhance mobility differences and to improve analyte separation. The authors hope that these discussions on derivatization and separation behaviors in nonaqueous media will significantly extend the application area of NACE with LIF detection.
Acknowledgement The authors express their sincere gratitude to Dr. Zhiyong Tang for helpful discussions about nonaqueous solvent chemistry.
references
[1] J. Kraly, M.A. Fazal, R.M. Schoenherr, R. Bonn, M.M. Harwood, E. Turner, M. Jones, N.J. Dovichi, Anal. Chem. 78 (2006) 4097. [2] M. Lacroix, V. Poinsot, C. Fournier, F. Couderc, Electrophoresis 26 (2005) 2608.
[3] Y.-W. Lin, T.-C. Chiu, H.-T. Chang, J. Chromatogr. B 793 (2003) 37. [4] V.L. Ward, M.G. Khaledi, J. Chromatogr. B 718 (1998) 15. [5] C. Massip, P. Riollet, E. Quemener, C. Bayle, R. Salvayre, F. ´ J. Chromatogr. A 979 (2002) 209. Couderc, E. Causse, [6] B. Valeur, Molecular Fluorescence: Principles and Applications, Wiley-VCH, Weinheim, 2001. [7] M. Molina, M. Silva, Electrophoresis 23 (2002) 3907. [8] A. Marsh, B. Clark, M. Broderick, J. Power, S. Donegan, K. Altria, Electrophoresis 25 (2004) 3970. [9] G.G. Guilbault (Ed.), Practical Fluorescence, Marcel Dekker, New York, 1990. [10] S. Hu, P.C.H. Li, J. Chromatogr. A 876 (2000) 183. [11] I. Beijersten, D. Westerlund, J. Chromatogr. A 716 (1995) 389. [12] S. Descroix, I.L. Potier, C. Niquet, N. Minc, J. Viovy, M. Taverna, J. Chromatogr. A 1087 (2005) 203. [13] J. Xie, J. Zhang, H. Liu, J. Liu, J. Tian, X. Chen, Z. Hu, Biomed. Chromatogr. 18 (2004) 600. [14] J. Zhang, J. Xie, J. Liu, J. Tian, X. Chen, Z. Hu, Electrophoresis 25 (2004) 74. ¨ [15] C. Huhn, M. Putz, N. Martin, R. Dahlenburg, U. Pyell, Electrophoresis 26 (2005) 2391. [16] E.M.R. Rodr´ıguez, M.S. Alaejos, C.D. Romero, Anal. Chim. Acta 334 (1996) 161. [17] X. Zhu, P.N. Shaw, D.A. Barrett, Anal. Chim. Acta 478 (2003) 259. [18] J.A. Monti, S.T. Christian, W.A. Shaw, J. Lipid Res. 19 (1978) 222. [19] M. Bem, F. Badea, C. Draghici, M.T. Caproiu, M. Vasilescu, M. Voicescu, A. Beteringhe, A. Caragheorgheopol, M. Maganu, T. Constantinescu, A.T. Balaban, ARKIVOC xiii (2007) 87. [20] S. Uchiyama, T. Santa, N. Okiyama, T. Fukushima, K. Imai, Biomed. Chromatogr. 15 (2001) 295. [21] L. Zhou, S. Wang, K. Tian, Y. Dong, Z. Hu, J. Sep. Sci. 30 (2007) 110. ¨ [22] H. Watzig, M. Degenhardt, A. Kunkel, Electrophoresis 19 (1998) 2695. [23] W.J. Lambert, D.L. Middleton, Anal. Chem. 62 (1990) 1585. [24] Y. Dong, X. Chen, Z. Hu, Microchim. Acta 154 (2006) 281. [25] Y. Dong, Y. Chen, X. Chen, Z. Hu, Biomed. Chromatogr. 20 (2006) 1150. ´ J. Chromatogr. [26] M.L. Riekkola, M. Jussila, S.P. Porras, I.E. Valko, A 892 (2000) 155. [27] F. Steiner, M. Hassel, Electrophoresis 21 (2000) 3994. [28] S. Hoffstetter-Kuhn, A. Paulus, E. Gassmann, H.M. Widmer, Anal. Chem. 63 (1991) 1541. [29] J.P. Landers, R.P. Oda, M.D. Schuchard, Anal. Chem. 64 (1992) 2846. [30] M. Stefansson, M. Novotny, J. Am. Chem. Soc. 115 (1993) 11573. [31] W. Lu, G.K. Poon, P.L. Carmichael, R.B. Cole, Anal. Chem. 68 (1996) 668. [32] A. Cifuentes, J.L. Bernal, J.C. Diez-Masa, Anal. Chem. 69 (1997) 4271.
a n a l y t i c a c h i m i c a a c t a 6 1 1 ( 2 0 0 8 ) 212–219
[33] R.M. Seifar, J.C. Kraak, W.T. Kok, Anal. Chem. 69 (1997) 2772. [34] R. Carabias-Mart´ınez, E. Rodr´ıguez-Gonzalo, J. ´ ´ Dom´ınguez-Alvarez, J. Hernandez-M endez, Anal. Chem. 69 (1997) 4437.
219
[35] J. Zhang, X. Chen, Z. Hu, X. Ma, Anal. Chim. Acta 471 (2002) 203. [36] M. Du, V. Flanigan, Y. Ma, Electrophoresis 25 (2004) 1496.