Heart-Cutting Two-Dimensional Capillary Electrophoresis for the On

Heart-cutting two-dimensional (2D) capillary electrophoresis (CE) in a single capillary was used for analysis of derivatized amino acids. A mixture of...
0 downloads 0 Views 308KB Size
Download PDF
Anal. Chem. 2008, 80, 1730-1736

Heart-Cutting Two-Dimensional Capillary Electrophoresis for the On-Line Purification and Separation of Derivatized Amino Acids Suzanne Anouti,† Odile Vandenabeele-Trambouze,† Dusˇan Koval,‡ and Herve´ Cottet*,†

Institut des Biomole´ cules Max Mousseron (UMR 5247 CNRSsUniversite´ de Montpellier 1sUniversite´ de Montpellier 2), place Euge` ne Bataillon CC 017, 34095 Montpellier Cedex 5, France, and UÄ stav Organicke´ Chemie a Biochemie AV C ˇ R, v.v.i., Flemingovo na´ m. 2. 166 10 Prague 6, Czech Republic

Heart-cutting two-dimensional (2D) capillary electrophoresis (CE) in a single capillary was used for analysis of derivatized amino acids. A mixture of 12 amino acids derivatized with UV-active benzyl 4-(3-(2-chloroethyl)-3nitrosoureido)butylcarbamate label served as a model of a moderately complex sample due to the presence of numerous derivatization byproducts. The first step of the heart-cutting 2D approach was sample cleanup by capillary zone electrophoresis (CZE) in borate electrolyte. Then, only a selected portion of the first-dimension separation was transferred into the second dimension of the separation by a specific voltage and pressure program. Finally, the zone of derivatized amino acids was separated by micellar electrokinetic chromatography in a boratesodium dodecyl sulfate system. The whole 2D process can be performed in a conventional CE analyzer without any interface for connection of the two separation modes. Intraday repeatability of the total migration time was 2%. In general, the heart-cutting 2D-CE methodology in a single capillary can be adapted for any CE mode regardless of the direction and velocity of electroosmotic flow and position of the fraction of interest in the first dimension (i.e., first, last, or intermediate fraction). For complex mixtures, the peak capacity of separation techniques such as high-performance liquid chromatography (HPLC), capillary electrophoresis (CE), and capillary electrochromatography (CEC) is insufficient for complete separation. Two-dimensional (2D) separation is a common approach to increase the separation power since peak capacity of the comprehensive 2D separation is given by the product of the individual separation peak capacities assuming that the separation mechanisms are orthogonal (i.e., are based on different molecular characteristics).1 2D-CE as a challenging issue from a technical point of view has attracted several research groups during the past decade. Dovichi’s group developed an interface with aligned capillaries * Corresponding author. Phone: +33-4-6714-3427. Fax: +33-4-6763-1046. E-mail: [email protected]. † UMR 5247 CNRSsUniversite´ de Montpellier 1sUniversite´ de Montpellier 2. ‡ U Ä stav Organicke´ Chemie a Biochemie AV C ˇ R. (1) Giddings, J. C. Unified Separation Science; John Wiley and Sons: New York, 1991.

1730 Analytical Chemistry, Vol. 80, No. 5, March 1, 2008

for proteomic analysis by 2D-CE.2-6 Their recent setup combines capillary sieving electrophoresis (CSE) with micellar electrokinetic chromatography (MEKC) and postcolumn fluorescence detection.2,3 Further, a microdialysis junction was designed by Zhang’s group to couple capillary isoelectric focusing (CIEF) with CSE.7,8 A similar interface was used by Mohan and Lee for analysis of protein tryptic digest by CIEFscapillary zone electrophoresis (CZE) coupling with a transient isotachophoresis step between the two dimensions.9 Recently, Sahlin published a CZE × MEKC separation of peptides in a system based on tangentially connected capillaries.10 Proteomic mixtures were subjected to CIEF × LC by Chen et al.11 and CIEF × CEC by Zhang and El Rassi,12 both using commercial microinjectors for connecting the two separation modes. In microfluidic format, Ramsey’s group reported a combination of MEKC and fast zone electrophoresis (ZE)13,14 and electrochromatography and ZE15 for analysis of protein tryptic digest. In a previous work, we demonstrated that it is alternatively possible to implement heart-cutting 2D-CE in a single capillary.16 In heart-cutting 2D separations, the second dimension of the (2) Zhu, C. R.; He, X. Y.; Kraly, J. R.; Jones, M. R.; Whitmore, C. D.; Gomez, D. G.; Eggertson, M.; Quigley, W.; Boardman, A.; Dovichi, N. J. Anal. Chem. 2007, 79, 765-768. (3) Kraly, J. R.; Jones, M. R.; Gomez, D. G.; Dickerson, J. A.; Harwood, M. M.; Eggertson, M.; Paulson, T. G.; Sanchez, C. A.; Odze, R.; Feng, Z. D.; Reid, B. J.; Dovichi, N. J. Anal. Chem. 2006, 78, 5977-5986. (4) Michels, D. A.; Hu, S.; Dambrowitz, K. A.; Eggertson, M. J.; Lauterbach, K.; Dovichi, N. J. Electrophoresis 2004, 25, 3098-3105. (5) Hu, S.; Michels, D. A.; Fazal, M. A.; Ratisoontorn, C.; Cunningham, M. L.; Dovichi, N. J. Anal. Chem. 2004, 76, 4044-4049. (6) Michels, D. A.; Hu, S.; Schoenherr, R. M.; Eggertson, M. J.; Dovichi, N. J. Mol. Cell. Proteomics 2002, 1, 69-74. (7) Yang, C.; Liu, H. C.; Yang, Q.; Zhang, L. Y.; Zhang, W. B.; Zhang, Y. K. Anal. Chem. 2003, 75, 215-218. (8) Liu, H. C.; Yang, C.; Yang, Q.; Zhang, W. B.; Zhang, Y. K. J. Chromatogr., B 2005, 817, 119-126. (9) Mohan, D.; Lee, C. S. Electrophoresis 2002, 23, 3160-3167. (10) Sahlin, E. J. Chromatogr., A 2007, 1154, 454-459. (11) Chen, J. Z.; Lee, C. S.; Shen, Y. F.; Smith, R. D.; Baehrecke, E. H. Electrophoresis 2002, 23, 3143-3148. (12) Zhang, M. Q.; El Rassi, Z. J. Proteome Res. 2006, 5, 2001-2008. (13) Rocklin, R. D.; Ramsey, R. S.; Ramsey, J. M. Anal. Chem. 2000, 72, 52445249. (14) Ramsey, J. D.; Jacobson, S. C.; Culbertson, C. T.; Ramsey, J. M. Anal. Chem. 2003, 75, 3758-3764. (15) Gottschlich, N.; Jacobson, S. C.; Culbertson, C. T.; Ramsey, J. M. Anal. Chem. 2001, 73, 2669-2674. (16) Cottet, H.; Biron, J. P.; Taillades, J. J. Chromatogr., A 2004, 1051, 25-32. 10.1021/ac702117h CCC: $40.75

© 2008 American Chemical Society Published on Web 01/30/2008

separation is only performed on one selected fraction coming from the first dimension. On the other hand, comprehensive 2D separations refer to separations wherein all the fractions coming from the first dimension are submitted to the second dimension. Nevertheless, in many analytical issues, the basic objective is the separation of one (or few) compound(s) in a complex mixture. In these cases, the heart-cutting 2D methodology is well adapted. To perform heart-cutting 2D-CE in a single capillary, a fraction of interest, stemming from the first dimension of the separation, is selected and isolated in the capillary by evacuating out of the capillary the other undesirable compounds. Next, the isolated fraction is separated in a second separation medium that is introduced in the same capillary by electroosmotic flow (EOF).16 The second separation medium is able to reach the isolated fraction since the solutes migrate in counter-EOF mode. This methodology is, however, limited to the separation of solutes migrating in counter-EOF mode and to the use of a seconddimension electrolyte that can enter in the capillary with an apparent velocity higher or identical to the electroosmotic velocity.16 In the present work, we report the possibility to extend the heart-cutting 2D-CE methodology in single capillary without any restriction on the composition of the electrolyte of the second dimension. Moreover, it is demonstrated that this 2D-CE methodology can be used for the on-line purification (first dimension) and separation (second dimension) of complex mixtures. Indeed, the detection of amino acids generally requires their derivatization for UV or fluorescence detection.17-20 A major analytical issue is related to the excess of derivatization reagent and/or the presence of undesirable compounds coming from secondary reactions occurring during the derivatization step21-23 or from impurities present in the derivatizing reagent.24-27 These undesirable compounds generally interfere with the derivatized amino acids in the separation. Therefore, one of the objectives of this work is to implement heart-cutting 2D-CE for on-line sample cleanup (first dimension) and separation (second dimension) of derivatized amino acids. EXPERIMENTAL SECTION Apparatus and Capillaries. Capillary electrophoresis was carried out with a PACE MDQ Beckman Coulter (Fullerton, CA) apparatus. Bare fused-silica capillaries of 25 or 50 µm i.d. and 360 µm o.d. were from Polymicro Technologies (Phoenix, AZ). Total length of the capillary was 30 cm with 20 cm distance from the (17) Poinsot, W.; Lacroix, M.; Maury, D.; Chataigne, G.; Feurer, B.; Couderc, F. Electrophoresis 2006, 27, 176-194. (18) Poinsot, V.; Bayle, C.; Couderc, F. Electrophoresis 2003, 24, 4047-4062. (19) Prata, C.; Bonnafous, P.; Fraysse, N.; Treilhou, M.; Poinsot, V.; Couderc, F. Electrophoresis 2001, 22, 4129-4138. (20) Smith, J. T. Electrophoresis 1999, 20, 3078-3083. (21) Frazier, R. A.; Papadopoulou, A. Electrophoresis 2003, 24, 4095-4105. (22) Underberg, W. J. M.; Waterval, J. C. M. Electrophoresis 2002, 23, 39223933. (23) Bardelmeijer, H. A.; Lingeman, H.; Deruiter, C.; Underberg, W. J. M. J. Chromatogr., A 1998, 807, 3-26. (24) Kang, X. J.; Xiao, J.; Huang, X.; Gu, Z. Z. Clin. Chim. Acta 2006, 366, 352356. (25) You, J. M.; Ming, Y. F.; Shi, Y. W.; Zhao, X. N.; Suo, Y. R.; Wang, H. L.; Li, Y. L.; Sun, J. Talanta 2005, 68, 448-458. (26) Ummadi, M.; Weimer, B. C. J. Chromatogr., A 2002, 964, 243-253. (27) Matsunaga, H.; Santa, T.; Hagiwara, K.; Homma, H.; Imai, K.; Uzu, S.; Nakashima, K.; Akiyama, S. Anal. Chem. 1995, 67, 4276-4282.

Scheme 1. Derivatization of Amino Acids with Z-CENU Reagent and Example of a Secondary Product Formationa

a Z-CENU 1 is transformed at alkaline conditions to isocyanate 2 which reacts with amino acid 3 (R is the amino acid side chain) to form stable derivative 4 (route A); isocyanate 2 can decompose to amine 5 and further react with excess 2 to form symmetrical urea byproduct 6 (route B).

inlet end to the detector. New capillaries were conditioned by rinsing with 1 M NaOH for 20 min and background electrolyte (BGE) for 10 min at 50 psi (3.45 bar). Between two runs, the capillary was successively washed at 50 psi with water for 2 min, 1 M NaOH for 2 min, and BGE for 6 min (25 µm i.d.) or 4 min (50 µm i.d.). Samples were introduced hydrodynamically at 0.3 psi (20 mbar) for 4 s (25 µm i.d.) or 3 s (50 µm i.d.). The temperature was set at 25 °C, and UV absorbance was monitored at 214 nm. Chemicals. L-Glutamic acid (Glu), l-aspartic acid (Asp), L-proline (Pro), and L-valine (Val) were from Prolabo (Briare-leCanal, France). Sodium tetraborate decahydrate was from Fluka (Buchs, Germany). Sodium dodecyl sulfate (SDS), L-leucine (Leu), and L-alanine (Ala) were from Merck (Overijse, Belgium). Boric acid, L-serine (Ser), and glycine (Gly) were purchased from Avocado (Heysham, England). D,L-R-Methyleucine (Mleu), Lisovaline (Iva), and 2-aminoisobutyric acid (AIB) were supplied by Acros Organics (Geel, Belgium). Isoleucine (Ile) was from Janssen Chimica (Geel, Belgium). Deionized water purified with a Milli-Q system (Millipore, Molsheim, France) was used in all the experimental work. Preparation of Derivatized Amino Acids. The derivatized reagent benzyl 4-(3-(2-chloroethyl)-3-nitrosoureido)butylcarbamate (Z-CENU) was synthesized according to the procedure described elsewhere.28,29 Briefly, 0.985 g of 2-chloroethyl 4-nitrobenzoyl(nitroso)carbamate dissolved in 15 mL of DMF was introduced dropwise to 1.22 g of benzyloxycarbonyl-putrescine in 15 mL of DMF. After 3 h at room temperature, the mixture was concentrated by vacuum, and the resulting oil was dissolved in CH2Cl2 (28) Martinez, J.; Oiry, J.; Imbach, J. L.; Winternitz, F. J. Med. Chem. 1982, 25, 178-182. (29) Vandenabeele, O.; Garrelly, L.; Ghelfenstein, M.; Commeyras, A.; Mion, L. J. Chromatogr., A 1998, 795, 239-250.

Analytical Chemistry, Vol. 80, No. 5, March 1, 2008

1731

Scheme 2. Schematic Representation of the Different Steps in Heart-Cutting 2D-CEa

a (A) 2D-CE methodology applicable when the fraction of interest (fraction B) is the last fraction of the first dimension and when the second dimension of the separation can enter in the capillary by EOF pumping. (B) Generalization of the 2D-CE methodology without any restriction on the definition of the second electrolyte and on the position of the fraction of interest (fraction B) in the first dimension: (1) first-dimension separation (injection by the inlet end of the capillary); (2) isolation and return of the fraction of interest (fraction B) to the inlet end of the capillary; (3) starting of the second-dimension separation; (4) separation of the fraction of interest in the second dimension.

and washed with 1 M HCl and saturated NaHCO3 solution until disappearance of the yellow color of the aqueous phase (4nitrophenol, side product). The organic phase was dried by anhydrous Na2SO4 and concentrated to afford 1.1 g (90%) of Z-CENU. Derivatization was performed for each individual amino acid as follows: 5 mL of 1 mM amino acid solution in a water/acetone mixture (40/60 v/v) was introduced into a 10 mL glass vial, and the pH was adjusted at 10.5-10.8 with pure triethylamine (TEA) (∼6.5 µL). Next, 500 µL (10 equiv) of 100 mM Z-CENU solution in acetone was added and the vials were sonicated for 2 h. The derivatization blank was prepared from a water/acetone mixture (40/60 v/v) devoid of amino acid following the same procedure. The pH was adjusted to 10.5 with TEA (∼4.5 µL). A sample mixture of 12 derivatized amino acids (1 mM each) was prepared by mixing 1 mL of each derivatized amino acid solution in a 25 mL flask. After evaporation of the solvent, the resulting solid was dissolved in 1 mL of a water/acetone mixture (40/60 v/v). RESULTS AND DISCUSSION Extension of the Heart-Cutting 2D-CE Methodology. In our previous work,16 we described the possibility of performing heartcutting 2D-CE in a single capillary according to the four key steps depicted in Scheme 2A. This methodology was applied for the charge-based (first dimension) and the size-based (second dimension) separation of polyelectrolytes. The first dimension was performed in free solution, whereas the second was realized in entangled polymer solution. At the end of the first-dimension separation, the fraction of interest (fraction B) was isolated in the 1732

Analytical Chemistry, Vol. 80, No. 5, March 1, 2008

capillary by evacuating fraction A by the outlet end of the capillary (step 1). Next, fraction B reached back the inlet end of the capillary by applying the voltage of reversed polarity (step 2). The second-dimension separation medium (containing a neutral sieving polymer for separation according to molar mass) was introduced by EOF in steps 3 and 4. The second dimension of the separation was effective owing to the migration in counter-EOF mode (separation of anionic compounds in a fused-silica capillary with strong cathodic EOF). Therefore, the second-dimension BGE could enter into the capillary with an apparent velocity higher than those of the solutes. In this work, we propose to extend the heart-cutting 2D-CE methodology without any restriction on either the definition of the BGE of the second dimension, on the mode of migration (i.e., co-EOF, counter-EOF, or suppressed EOF), or on the position of the fraction of interest in the first dimension (i.e., the first, last, or intermediate fraction). Contrary to ref 16, the isolation of fraction B, described in Scheme 2B, step 2, is performed hydrodynamically instead of reversing the polarity of the voltage. Therefore, fractions B and C cannot recombine. This step 2 is also accompanied with the hydrodynamic introduction of the second-dimension BGE into the capillary. As a consequence, fraction B can be separated in the second medium whatever is the composition of the second-dimension BGE. Application to the On-Line Purification and Separation of Derivatized Amino Acid Complex Mixtures. In this work, 12 amino acids were derivatized using the Z-CENU reagent 1 (see Scheme 1, route A). This reagent possesses a protected isocyanate

Scheme 3. Principle of the Application of Heart-Cutting 2D-CE to the Purification (First Dimension) and the Separation (Second Dimension) of Complex Mixtures in a Single Capillarya

Figure 1. Electropherograms showing the monodimensional separation of 12 Z-CENU-derivatized amino acids in the presence (A) or in the absence (C) of SDS. Trace B displays the separation of the derivatization blank in the presence of SDS. Experimental conditions: fused-silica capillary, 30 cm (20 cm from the inlet end to the detector) × 50 µm i.d.; BGEs, 200 mM borate/100 mM sodium buffer (C) + 50 mM SDS (A and B) in a (85/15 v/v) water/isopropyl alcohol mixture, pH* 9.7 (A and B) and pH* 9.6 (C); voltage, +20 kV; hydrodynamic injection, 0.3 psi, 3 s; UV detection at 214 nm; temperature, 25 °C. Peak identification: 1, Pro; 2, Ser; 3, AIB; 4, Iva; 5, Ala; 6, Gly; 7, Mleu; 8, Val; 9, Ileu; 10, Leu; 11, Glu; 12, Asp; /, secondary compounds.

group and is suitable for the derivation of primary and secondary amines.29-31 Its modular structure and the derivative stability over time allows its use for different analytical techniques (HPLC, CE) as well as production of highly specific antibodies used in ELISA analysis.28,30,32,33 It was demonstrated that the derivatization with Z-CENU is a robust method allowing quantitative derivatization of amino compounds in a concentration range from 10-1 to 10-7 M.32 However, due to the Z-CENU excess under typical derivatization conditions, secondary compounds are produced such as neutral symmetrical urea 6 as depicted in Scheme 1, route B.29 Typically, amino acids derivatized by conventional labels are separated in MEKC due to hydrophobic interactions with the micellar pseudostationary phase.21 Accordingly, a mixture of 12 Z-CENU amino acids was resolved in a sodium borate-SDS BGE (50 mM SDS, 200 mM borate buffer pH* 9.7 in a (85/15 v/v) water/isopropyl alcohol) (see Figure 1A). Hydroorganic electrolyte was used for a better solubility of the Z-CENU-derivatized amino acids. However, a number of secondary compounds, noted with an asterisk in Figure 1, comigrated with the solutes of interest. As shown in Figure 1B for the derivatization blank at (30) Claeys-Bruno, A.; Vandenabeele-Trambouze, O.; Sergent, M.; Geffard, A.; Bodet, D.; Dobrijevic, A.; Commeyras, A.; Luu, R. P. T. Chemom. Intell. Lab. Syst. 2006, 80, 176-185. (31) Vandenabeele-Trambouze, O.; Geffard, M.; Bodet, D.; Despois, M.; Dobrijevic, M.; Loustalot, M. F. G.; Commeyras, A. Chirality 2002, 14, 519526. (32) Claeys-Bruno, M.; Vandenabeele-Trambouze, O.; Sergent, M.; Geffard, M.; Bodet, D.; Dobrijevic, M.; Commeyras, A.; Luu, R. P. T. Chemom. Intell. Lab. Syst. 2006, 80, 186-197. (33) Vandenabeele-Trambouze, O.; Albert, M.; Bayle, C.; Couderc, F.; Commeyras, A.; Despois, D.; Dobrijevic, M.; Loustalot, M. F. G. J. Chromatogr., A 2000, 894, 259-266.

a (1) Sample cleanup (evacuation of fraction A) in the first dimension (injection by the outlet end of the capillary); (2) return and recombination of the fractions of interest (fractions B and C) at the outlet end of the capillary by switching the polarity; (3) introduction of the second-dimension electrolyte (BGE2) by hydrodynamic flow; (4) separation of fractions B and C in the second dimension.

the same electrophoretic conditions, these undesirable compounds are directly related to secondary reactions involving the Z-CENU derivatization reagent. Although these compounds were not all identified, an example of symmetrical urea byproduct 6 is given in Scheme 1. Accordingly, these secondary compounds are neutral and comigrate with the EOF in the absence of SDS as shown in Figure 1C. Therefore, the issue is to purify the sample by evacuating these neutral species in a first dimension (nonmicellar mode) and, next, to separate the purified sample in a second dimension (micellar mode). Following the general concept described in the previous section, Scheme 3 depicts the four different steps designed for that purpose. The corresponding electropherogram obtained on a single 50 µm i.d. fused-silica capillary is given in Figure 2. In step 1, the derivatized amino acid sample is injected by the outlet end of the capillary and a nonmicellar separation is performed in the first dimension (BGE1, 200 mM borate buffer, pH* 9.6 in a water/isopropyl alcohol mixture (85/15 v/v), -20 kV). During this separation stage, the sample is separated into three main fractions. The first one (fraction A) corresponds to the undesirable neutral species that comigrate with EOF. The less mobile fractions (B and C) are related to the derivatized amino acids. At the end of step 1, the unwanted fraction A is evacuated out of the capillary Analytical Chemistry, Vol. 80, No. 5, March 1, 2008

1733

Figure 2. Heart-cutting 2D-CE in a single capillary for the purification and the separation of a mixture containing 12 Z-CENU-derivatized amino acids. Experimental conditions: fused-silica capillary, 30 cm (20 cm from the inlet end to the detector) × 50 µm i.d.; BGEs, first dimension (BGE1), 200 mM borate/100 mM sodium buffer, pH* 9.6 in a (85/15 v/v) water/isopropyl alcohol mixture, second dimension (BGE2), BGE1 + 50 mM SDS, pH* 9.7; voltage program (see plain line on the upper part of the graph), -20 kV, t ) 0-4.82 min; +20 kV, t ) 4.82-9.69 min; 0 kV, t ) 9.69-19.75 min; +20 kV, t > 19.75 min; pressure program (see dotted line on the graph), 0.4 psi at the outlet end of the capillary, t ) 9.69-19.75 min, no pressure when not specified; hydrodynamic injection by the outlet end, 0.3 psi, 3 s; detection, UV 214 nm; temperature, 25 °C. Peak identification is as in Figure 1.

by the inlet end. In step 2, the high-voltage polarity is reversed so that the purified derivatized amino acids (fractions B and C) pass again through the detection point and rejoin at the outlet end of the capillary. To avoid any loss in sample from the outlet end of the capillary at the end of step 2, a plug of BGE1 (0.3 psi for 9 s) is initially injected behind the sample at the beginning of step 1. In step 3, the outlet vial is changed with the seconddimension electrolyte (BGE2, depicted in gray in Scheme 3). The BGE2 enters the capillary by hydrodynamic flow (0.4 psi, 29 mbar) directed from the outlet to the inlet end of the capillary. Finally, in step 4, the purified sample is separated in micellar mode (BGE2, 200 mM borate buffer containing 50 mM SDS, pH* 9.7 in a water/ isopropyl alcohol mixture (85/15 v/v), +20 kV). As expected, the secondary compounds were not detected in Figure 2 demonstrating the interest of this approach. Intraday repeatability of migration time was 2% for a series of 10 repetitions (see Figure 3). Optimization of the Mobilization Pressure and the Internal Diameter of the Capillary. The use of a hydrodynamic flow in step 3 reduces the peak efficiency in the second dimension due to peak broadening by Taylor dispersion.34-36 The parabolic velocity profile in the capillary generates the peak dispersion which is strongly dependent on the capillary internal diameter, the mobilization pressure (or the linear velocity), and the diffusion coefficient of the solute.34-36 To reduce this peak broadening effect and to achieve better resolution of the separation in the second (34) Taylor, G. Proc. R. Soc. London, Ser. A 1953, 219, 186-203. (35) Aris, R. Proc. R. Soc. London, Ser. A 1956, 235, 67-77. (36) Grushka, E.; Kikta, E. J. J. Phys. Chem. 1974, 78, 2297-2301.

1734

Analytical Chemistry, Vol. 80, No. 5, March 1, 2008

dimension, the effect of the capillary internal diameter and of the mobilization pressure was considered. As described by the Taylor-Aris equation, the height equivalent to a theoretical plate H is a function of linear velocity u according to the following equation:

dc2 1 H ) 2D + u u 96D

(1)

where D is the molecular diffusion coefficient and dc is the capillary internal diameter. Figure 4 displays the variation of H as a function of u for the Z-CENU Gly derivative in 25 and 50 µm i.d. capillaries. These H values were experimentally obtained after injection and mobilization of the individual solute on a given capillary length. As expected, the H value decreases with decreasing internal diameter for a given linear velocity value. In Figures 2 and 3 (50 µm i.d.), the linear velocity of the hydrodynamic flow during the mobilization step was ∼4 × 10-4 m‚s-1 (P ) 0.4 psi). For the same linear velocity in 25 µm i.d. capillary (P ) 1.6 psi), the H value is decreased by a factor 2.8. Moreover, at low velocities and for such small molecules (D ∼ 3.3 × 10-10 m2‚s-1, calculated from the slope of the ascending branch in Figure 4), the H ) f(u) plot is much flatter on a 25 µm i.d. capillary. The gain in peak efficiency by using the 25 µm i.d. capillary in heart-cutting 2D-CE can be observed in Figure 5A as compared to Figure 2. It should be noted that the differences in migration times between the two figures are mainly due to the difference in voltage in the first dimension (20 kV in Figure 2 vs 13 kV in Figure 5A). From the numerical

Figure 3. Electropherograms showing the repeatability of the heartcutting 2D-CE for the purification and the separation 12 Z-CENUderivatized amino acids. Electrophoretic conditions were as in Figure 2. Three replications out of 10 are displayed.

Figure 4. Variation of the height equivalent to a theoretical plate, H, as a function of the linear velocity, u, of the hydrodynamic flow in 25 and 50 µm i.d. capillaries. Experimental data were obtained by injection and mobilization of 1 mM Z-CENU Glu derivative dissolved in the 200 mM borate/100 mM sodium buffer pH* 9.6 in the (85/15 v/v) water/isopropyl alcohol mixture. The diffusion coefficient (D ) 3.3 × 10-10 m2‚s-1) of the solute was obtained by curve fitting according to eq 1.

values on peak efficiency reported in Table 1 for Z-CENU Leu and Ile derivatives, the gain in peak efficiency for the 2D-CE varies from 1.3 to 2.6 and is in reasonable agreement with the maximal gain expected from the H ) f(u) curves. Decreasing further the capillary internal diameter to get better peak efficiency would be difficult to implement in practice due to low detection sensitivity. Moreover, there is a constant contribution to the H value (∼8 µm) that is due to the injected bandwidth and that cannot be decreased by using lower internal capillary diameters. As reported in previous work,16 25 µm i.d. seems to be a good compromise between detection sensitivity and peak efficiency in heart-cutting 2D-CE with UV detection. In fact, the plate heights for the 25 and

Figure 5. Effect of sample stacking on the purification and separation by heart-cutting 2D-CE of a mixture containing 12 Z-CENUderivatized amino acids in a 25 µm i.d. capillary. Experimental conditions: fused-silica capillary, 30 cm (20 cm from the inlet end to the detector) × 25 µm i.d.; BGEs, first dimension (BGE1), 200 mM borate/100 mM sodium buffer, pH* 9.6 (A), 50 mM borate/25 mM sodium buffer, pH* 9.6 (B), 50 mM borate/7 mM sodium pH 9.0 (C), all buffers were prepared in a water/isopropyl alcohol mixture (85/15 v/v); second dimension (BGE2), 200 mM borate/100 mM sodium buffer + 50 mM SDS, pH* 9.7 in a water/isopropyl alcohol mixture (85/15 v/v); voltage program, -13 kV for t ) 0-7.46 min (A), t ) 0-4.86 min (B), and t ) 0-3.26 min (C); +13 kV for t ) 7.46-14.97 min (A), t ) 4.86-9.77 min (B), t ) 3.26-6.53 min (C); 0 kV for t ) 14.97-24.47 min (A), t ) 9.77-19.57 min (B), t ) 6.53-15.53 min (C); +20 kV for t > 24.47 min (A), t > 19.57 min (B), t > 15.53 min (C); pressure program, 1.6 psi at the outlet end of the capillary, t ) 14.97-24.47 min (A), t ) 9.77-19.57 min (B), t ) 6.53-15.53 min (C); hydrodynamic injection by the outlet end, 0.3 psi, 3 s; UV detection at 214 nm; temperature, 25 °C. Peak identification is as in Figure 1.

50 µm capillaries at the pressures of 1.6 and 0.4 psi are about 17 and 47 µm, respectively (see Figure 4). These values correspond to a maximum possible plate numbers of about 12 000 and 4300, respectively (on 20 cm effective length), and are well below plate numbers given in Table 1. Presumably, zone sharpening due to a sweeping effect occurs in a transition between the first and second dimension of the separation as a consequence of the presence of micelles in BGE2. Increasing the Peak Efficiency by Field-Amplified Sample Stacking between the Two Dimensions. To overcome the peak broadening caused by the hydrodynamic flow and to achieve better separations, field-amplified sample stacking was carried out between the two dimensions of the separation. Sample stacking is based on the variation of the conductivity between the sample zone and the BGE. Consequently, stacking occurs due to the resulting difference in electric field (and velocities of the analyzed ions) at the interface between these two zones. For that purpose, the borate concentration of BGE1 was decreased from 200 mM to 50 mM at constant pH, keeping unchanged the conditions of the second dimension (BGE2, +20 kV). The effect of the sample stacking on the separations can be contemplated in Figure 5 (trace B vs trace A). The total analysis time decreased with decreasing ionic strength of BGE1 due to faster EOF in the first dimension. The starting point of the second dimension of the separation is Analytical Chemistry, Vol. 80, No. 5, March 1, 2008

1735

Table 1. Theoretical Plate Numbers Obtained by Heart-Cutting 2D-CE for Two Z-CENU Amino Acid Derivatives Using Different Capillary i.d.’s and BGEs in the First Dimension exptl conditions

capillary i.d. (µm)

borate concn BGE1/BGE2

pH* BGE1/BGE2

N (103) Z-CENU Ile

N (103) Z-CENU Leu

Figure 2 Figure 5A Figure 5B Figure 5C

50 25 25 25

200/200 200/200 50/200 50/200

9.6/9.7 9.6/9.7 9.6/9.7 8.0/9.7

18 27 43 56

25 28 47 57

labeled with an arrow in Figure 5. Differences in the spacing between peaks 1 and 12 in Figure 5, traces A and B, is due to a slight variation in the EOF. The field-amplified sample stacking effect was effective in reducing the peak broadening since the peak efficiency was increased by a factor of ∼1.6 (see lines 2 and 3 in Table 1, for Leu and Ile derivatives). As a consequence, better resolutions for the pairs 4/5 and 7/8 could be obtained. Finally, to further increase the difference in conductivity between the two electrolytes, the pH of the electrolyte used in the first dimension was decreased (50 mM boric acid in the (85/15 v/v) water/ isopropyl alcohol mixture, pH* 9.0). The separation is displayed in Figure 5 (trace C). Because of the low ionic strength of BGE1, EOF was increased in the first dimension and the total analysis time was shortened. The peak efficiency was increased by a factor of 1.2-1.3 as compared to Figure 5B (see the two last lines in Table 1) demonstrating that lowering the conductivity of BGE1 improved the sample stacking effect. CONCLUSION The use of 25 µm i.d. capillaries combined with sample stacking effects improved the performance of the heart-cutting 2D-CE for the on-line sample purification (first dimension) and MEKC separation of a derivatized amino acid mixture. This work also demonstrates the possibility to extend the heart-cutting 2D-CE methodology in a single capillary without any restriction on either the definition of the electrolyte of the second dimension, on the

1736 Analytical Chemistry, Vol. 80, No. 5, March 1, 2008

mode of migration (co-EOF, counter-EOF, or in absence of EOF), or on the position of the fraction of interest in the first dimension (first, last, or intermediate fraction). This opens up the possibility to combine electrolytes of virtually any composition in 2D heartcutting separations. In particular, it is now possible to consider the combination of electrolytes varying in pH, ionic strength, surfactant concentration, or chiral selector content. An important feature of this approach is that the heart-cutting 2D-CE can be performed on commercial CE apparatus without any hardware modification, as far as the control of the mobilization pressure is accurate. The use of a more sensitive detection mode, such as laser-induced fluorescence, should also favor the implementation of heart-cutting 2D-CE separation using capillaries with low internal diameters (e25 µm). ACKNOWLEDGMENT We gratefully acknowledge the support from the French Research National Agency (ACI jeunes chercheurs no. 4093) for the 2D-CE methodology.

Received for review December 18, 2007. AC702117H

October

15,

2007.

Accepted